refactor(agent): let runner own context compaction (#5568)

* refactor(agent): consolidate accepted history under pressure

* fix(agent): align provider and session compaction

* refactor(agent): simplify runner context compaction

* refactor(agent): remove background token consolidation

* fix(agent): keep injected transcript messages distinct

* refactor(agent): unify native compaction summaries

* fix(agent): preserve native compaction boundary

* fix(agent): unify context compaction paths

* fix(agent): preserve exact compaction request boundaries

nanobot/agent/context.py

@@ -235,7 +235,7 @@ class ContextBuilder:
     def build_messages(
         self,
         history: list[dict[str, Any]],
-        current_message: str,
+        current_message: str | None,
         *,
         media: list[str] | None = None,
         channel: str | None = None,
@@ -259,6 +259,8 @@ class ContextBuilder:
             workspace=workspace,
             include_memory=include_memory,
         )
+        if current_message is None:
+            return messages
         current = messages[-1]
         if len(messages) < 2 or messages[-2].get("role") != current.get("role"):
             return messages

nanobot/agent/context_governance.py

@@ -1,19 +1,43 @@
-"""Model-message governance for agent runner requests.
+"""Model-message governance and compaction for agent runner requests.
 
-This module owns model-facing message shaping and tool-result content normalization.
-It may return copied messages or persisted-result placeholders, but it must not
-mutate an existing session history list in place.
+This module owns model-facing message shaping, request pressure, H/delta
+compaction state, and tool-result content normalization. It may return copied
+messages or persisted-result placeholders, but it must not mutate an existing
+session history list in place.
 """
 
 from __future__ import annotations
 
-from dataclasses import dataclass
+from collections.abc import Awaitable, Callable
+from copy import deepcopy
+from dataclasses import dataclass, replace
+from datetime import datetime
 from pathlib import Path
 from typing import TYPE_CHECKING, Any, cast
 
 from loguru import logger
 
-from nanobot.providers.base import LLMUsage
+from nanobot.agent.context import TranscriptInput
+from nanobot.providers.base import (
+    LLMResponse,
+    LLMUsage,
+    ProviderCallContext,
+    ProviderConversationState,
+)
+from nanobot.providers.conversation_state import (
+    ProviderConversationStateController,
+    allows_conversation_message_merge,
+)
+from nanobot.runtime_context import (
+    RUNTIME_CONTEXT_MESSAGE_META,
+    detach_runtime_context,
+    reattach_runtime_context,
+)
+from nanobot.session.history_visibility import is_hidden_history_message
+from nanobot.session.summary import (
+    SUMMARY_CONTINUATION_TEXT,
+    SessionSummaryCheckpoint,
+)
 from nanobot.utils.helpers import (
     estimate_message_tokens,
     estimate_prompt_tokens_chain,
@@ -27,6 +51,16 @@ if TYPE_CHECKING:
     from nanobot.agent.tools.registry import ToolRegistry
     from nanobot.providers.base import LLMProvider
 
+TranscriptBuilder = Callable[[TranscriptInput], list[dict[str, Any]]]
+HistoryConsolidator = Callable[
+    [list[dict[str, Any]], str | None],
+    Awaitable[str | None],
+]
+ProviderCompactionConsolidator = Callable[
+    [ProviderConversationState, list[dict[str, Any]], str | None],
+    Awaitable[str | None],
+]
+
 SNIP_SAFETY_BUFFER = 1024
 # read_file is the recovery path for persisted results; exempting it prevents persist->read->persist loops.
 TOOL_RESULT_OFFLOAD_EXEMPT_TOOLS = frozenset({"read_file"})
@@ -85,8 +119,204 @@ class ContextGovernanceConfig:
     max_tokens: int | None = None
 
 
+@dataclass(slots=True)
+class ContextCompactionState:
+    """Track accepted provider input H separately from the unsent delta."""
+
+    raw_messages: list[dict[str, Any]]
+    accepted_messages: list[dict[str, Any]]
+    raw_accepted_boundary: int
+    active_summary: str | None
+    transcript_input: TranscriptInput
+    transcript_builder: TranscriptBuilder
+    consolidate_history: HistoryConsolidator
+    consolidate_provider_compaction: ProviderCompactionConsolidator | None
+    summary_checkpoint: SessionSummaryCheckpoint | None = None
+
+    @classmethod
+    def from_transcript(
+        cls,
+        transcript_input: TranscriptInput,
+        transcript_builder: TranscriptBuilder,
+        consolidate_history: HistoryConsolidator | None,
+        consolidate_provider_compaction: ProviderCompactionConsolidator | None,
+    ) -> tuple[list[dict[str, Any]], ContextCompactionState | None]:
+        """Build the raw transcript and its initial H/delta boundary."""
+        messages = list(transcript_builder(transcript_input))
+        if consolidate_history is None:
+            return messages, None
+        accepted_history_boundary = 1 + len(transcript_input.history)
+        return messages, cls(
+            raw_messages=messages,
+            accepted_messages=deepcopy(messages[:accepted_history_boundary]),
+            raw_accepted_boundary=accepted_history_boundary,
+            active_summary=(
+                transcript_input.session_summary["text"]
+                if transcript_input.session_summary is not None
+                else None
+            ),
+            transcript_input=transcript_input,
+            transcript_builder=transcript_builder,
+            consolidate_history=consolidate_history,
+            consolidate_provider_compaction=consolidate_provider_compaction,
+        )
+
+    def request_messages(
+        self,
+        raw_messages: list[dict[str, Any]],
+    ) -> list[dict[str, Any]]:
+        return [
+            *deepcopy(self.accepted_messages),
+            *deepcopy(raw_messages[self.raw_accepted_boundary:]),
+        ]
+
+    def delta_after_accepted(
+        self,
+        request_messages: list[dict[str, Any]],
+    ) -> list[dict[str, Any]]:
+        return deepcopy(request_messages[len(self.accepted_messages):])
+
+    def accept_request(
+        self,
+        model_messages: list[dict[str, Any]],
+        *,
+        raw_boundary: int,
+    ) -> None:
+        """Advance H after the provider has received one request."""
+        self.accepted_messages = deepcopy(model_messages)
+        self.raw_accepted_boundary = raw_boundary
+
+
+@dataclass(slots=True)
+class ModelRequestState:
+    """Context state shared by every provider request in one runner turn."""
+
+    config: ContextGovernanceConfig
+    conversation: ProviderConversationStateController
+    usage: LLMUsage | None = None
+    messages: list[dict[str, Any]] | None = None
+    tool_definitions: list[dict[str, Any]] | None = None
+    compaction: ContextCompactionState | None = None
+    provider_compaction_applied: bool = False
+
+
 class ContextGovernor:
-    """Prepare model-copy messages while preserving persisted history."""
+    """Own model-request context while preserving persisted history."""
+
+    @staticmethod
+    def _merge_message_content(left: Any, right: Any) -> str | list[dict[str, Any]]:
+        if isinstance(left, str) and isinstance(right, str):
+            return f"{left}\n\n{right}" if left else right
+
+        def _to_blocks(value: Any) -> list[dict[str, Any]]:
+            if isinstance(value, list):
+                return [
+                    cast(dict[str, Any], item)
+                    if isinstance(item, dict)
+                    else {"type": "text", "text": str(item)}
+                    for item in cast(list[Any], value)
+                ]
+            if value is None:
+                return []
+            return [{"type": "text", "text": str(value)}]
+
+        return _to_blocks(left) + _to_blocks(right)
+
+    @classmethod
+    def _merge_adjacent_user_messages_for_model(
+        cls,
+        messages: list[dict[str, Any]],
+    ) -> list[dict[str, Any]]:
+        """Merge adjacent visible user messages only in the model-facing copy."""
+        prepared: list[dict[str, Any]] = []
+        for source in messages:
+            injection = deepcopy(source)
+            if (
+                prepared
+                and injection.get("role") == "user"
+                and prepared[-1].get("role") == "user"
+                and injection.get("content") != SUMMARY_CONTINUATION_TEXT
+                and prepared[-1].get("content") != SUMMARY_CONTINUATION_TEXT
+                and not is_hidden_history_message(injection)
+                and not is_hidden_history_message(prepared[-1])
+                and allows_conversation_message_merge(injection)
+                and allows_conversation_message_merge(prepared[-1])
+            ):
+                merged = dict(prepared[-1])
+                left_meta = merged.get("_meta")
+                right_meta = injection.get("_meta")
+                left_meta_dict = (
+                    cast(dict[str, Any], left_meta) if isinstance(left_meta, dict) else None
+                )
+                right_meta_dict = (
+                    cast(dict[str, Any], right_meta) if isinstance(right_meta, dict) else None
+                )
+                left_marker = (
+                    left_meta_dict.get(RUNTIME_CONTEXT_MESSAGE_META)
+                    if left_meta_dict is not None
+                    else None
+                )
+                right_marker = (
+                    right_meta_dict.get(RUNTIME_CONTEXT_MESSAGE_META)
+                    if right_meta_dict is not None
+                    else None
+                )
+                left_marker_dict = (
+                    cast(dict[str, Any], left_marker) if isinstance(left_marker, dict) else None
+                )
+                right_marker_dict = (
+                    cast(dict[str, Any], right_marker) if isinstance(right_marker, dict) else None
+                )
+                empty_sources: list[str] = []
+                empty_blocks: list[dict[str, Any]] = []
+                detached_left = (
+                    detach_runtime_context(merged.get("content"), left_marker_dict)
+                    if left_marker_dict is not None
+                    else (merged.get("content"), empty_sources, empty_blocks)
+                )
+                detached_right = (
+                    detach_runtime_context(injection.get("content"), right_marker_dict)
+                    if right_marker_dict is not None
+                    else (injection.get("content"), empty_sources, empty_blocks)
+                )
+                if detached_left is not None and detached_right is not None:
+                    left_content, left_sources, left_blocks = detached_left
+                    right_content, right_sources, right_blocks = detached_right
+                    merged_content = cls._merge_message_content(left_content, right_content)
+                    context_blocks = [*left_blocks, *right_blocks]
+                    if context_blocks:
+                        merged_content, marker = reattach_runtime_context(
+                            merged_content,
+                            [*left_sources, *right_sources],
+                            context_blocks,
+                        )
+                        internal_meta = (
+                            dict(left_meta_dict) if left_meta_dict is not None else {}
+                        )
+                        if right_meta_dict is not None:
+                            for key, value in right_meta_dict.items():
+                                internal_meta.setdefault(key, value)
+                        internal_meta[RUNTIME_CONTEXT_MESSAGE_META] = marker
+                        merged["_meta"] = internal_meta
+                    merged["content"] = merged_content
+                else:
+                    merged["content"] = cls._merge_message_content(
+                        merged.get("content"),
+                        injection.get("content"),
+                    )
+                prepared[-1] = merged
+                continue
+            prepared.append(injection)
+        return prepared
+
+    def prepare_messages_for_model(
+        self,
+        config: ContextGovernanceConfig,
+        messages: list[dict[str, Any]],
+    ) -> list[dict[str, Any]]:
+        """Build the normalized model-facing copy of a raw transcript."""
+        governed = self.prepare_for_model(config, messages)
+        return self._merge_adjacent_user_messages_for_model(governed)
 
     def prepare_for_model(
         self,
@@ -115,17 +345,31 @@ class ContextGovernor:
         )
         updated = self.drop_orphan_tool_results(updated)
         updated = self.backfill_missing_tool_results(updated)
+        return self.ensure_request_fits(
+            config,
+            updated,
+            tool_definitions=tool_definitions,
+        )
+
+    def ensure_request_fits(
+        self,
+        config: ContextGovernanceConfig,
+        messages: list[dict[str, Any]],
+        *,
+        tool_definitions: list[dict[str, Any]] | None,
+    ) -> list[dict[str, Any]]:
+        """Validate an exact model request without dropping any messages."""
         if not config.context_window_tokens:
-            return updated
+            return messages
         budget = self.input_budget(config)
         estimated, source = estimate_prompt_tokens_chain(
             config.provider,
             config.model,
-            updated,
+            messages,
             tool_definitions,
         )
         if budget > 0 and estimated <= budget:
-            return updated
+            return messages
         raise ContextWindowExceededError(
             session_key=config.session_key,
             estimated_tokens=estimated,
@@ -133,6 +377,41 @@ class ContextGovernor:
             source=source,
         )
 
+    def request_pressure(
+        self,
+        config: ContextGovernanceConfig,
+        messages: list[dict[str, Any]],
+        usage: LLMUsage | None,
+        *,
+        usage_matches_messages: bool,
+        tool_definitions: list[dict[str, Any]] | None,
+        request_context_tokens: int | None = None,
+    ) -> tuple[int, str] | None:
+        """Return the authoritative measurement when a request is pressured."""
+        if not config.context_window_tokens:
+            return None
+        budget = self.input_budget(config)
+        if request_context_tokens is not None:
+            measured = request_context_tokens
+            source = "resumed provider state plus pending messages"
+        elif (
+            usage_matches_messages
+            and usage is not None
+            and usage.context_tokens is not None
+        ):
+            measured = usage.context_tokens
+            source = "matching provider usage"
+        else:
+            measured, source = estimate_prompt_tokens_chain(
+                config.provider,
+                config.model,
+                messages,
+                tool_definitions,
+            )
+        if budget > 0 and measured < budget:
+            return None
+        return measured, source
+
     def fit_request(
         self,
         config: ContextGovernanceConfig,
@@ -144,27 +423,15 @@ class ContextGovernor:
         request_context_tokens: int | None = None,
     ) -> tuple[list[dict[str, Any]], bool]:
         """Fit the request when its measured or estimated input is pressured."""
-        if not config.context_window_tokens:
-            return messages, False
-        budget = self.input_budget(config)
-        if (
-            request_context_tokens is None
-            and usage_matches_messages
-            and usage is not None
-            and usage.context_tokens is not None
-        ):
-            pressured = budget <= 0 or usage.context_tokens >= budget
-        else:
-            estimated, _ = estimate_prompt_tokens_chain(
-                config.provider,
-                config.model,
-                messages,
-                tool_definitions,
-            )
-            if request_context_tokens is not None:
-                estimated = max(estimated, request_context_tokens)
-            pressured = budget <= 0 or estimated >= budget
-        if not pressured:
+        pressure = self.request_pressure(
+            config,
+            messages,
+            usage,
+            usage_matches_messages=usage_matches_messages,
+            tool_definitions=tool_definitions,
+            request_context_tokens=request_context_tokens,
+        )
+        if pressure is None:
             return messages, False
         return self.fit_to_budget(
             config,
@@ -172,6 +439,201 @@ class ContextGovernor:
             tool_definitions=tool_definitions,
         ), True
 
+    @staticmethod
+    def _summary_transcript(
+        compaction: ContextCompactionState,
+        summary: str,
+    ) -> list[dict[str, Any]]:
+        """Rebuild only the stable system prefix around a replacement summary."""
+        return compaction.transcript_builder(
+            replace(
+                compaction.transcript_input,
+                history=[],
+                current_message=None,
+                media=None,
+                session_summary={
+                    "text": summary,
+                    "last_active": datetime.now().astimezone().isoformat(),
+                },
+                runtime_context_blocks=None,
+            )
+        )
+
+    async def summarize_provider_compaction(
+        self,
+        state: ModelRequestState,
+        response: LLMResponse,
+        *,
+        current_request_boundary: int | None,
+    ) -> None:
+        """Materialize the exact input replaced by provider-native compaction."""
+        compaction = state.compaction
+        if (
+            not response.provider_compaction_applied
+            or response.provider_compaction_state is None
+            or compaction is None
+            or compaction.consolidate_provider_compaction is None
+        ):
+            return
+
+        if response.provider_compaction_scope == "prior_context":
+            accepted_messages = compaction.accepted_messages
+            transcript_boundary = compaction.raw_accepted_boundary
+        elif (
+            response.provider_compaction_scope == "current_request"
+            and state.messages is not None
+            and current_request_boundary is not None
+        ):
+            accepted_messages = state.messages
+            transcript_boundary = current_request_boundary
+        else:
+            logger.warning(
+                "Ignoring provider compaction with missing request-boundary scope for {}",
+                state.config.session_key or "default",
+            )
+            return
+
+        summary = await compaction.consolidate_provider_compaction(
+            response.provider_compaction_state,
+            deepcopy(accepted_messages),
+            compaction.active_summary,
+        )
+        if not summary:
+            return
+        compaction.active_summary = summary
+        compaction.summary_checkpoint = SessionSummaryCheckpoint(
+            summary=summary,
+            transcript_boundary=transcript_boundary,
+        )
+
+    async def _compact_request_history(
+        self,
+        state: ModelRequestState,
+        compaction: ContextCompactionState,
+        messages: list[dict[str, Any]],
+        pressure: tuple[int, str],
+        *,
+        tool_definitions: list[dict[str, Any]] | None,
+    ) -> list[dict[str, Any]]:
+        """Replace accepted history H with a checkpoint while preserving delta."""
+        delta_messages = compaction.delta_after_accepted(messages)
+        consolidation_prefix = self.prepare_messages_for_model(
+            state.config,
+            compaction.accepted_messages,
+        )
+        summary = await compaction.consolidate_history(
+            deepcopy(consolidation_prefix),
+            compaction.active_summary,
+        )
+        if not summary:
+            measured, source = pressure
+            raise ContextWindowExceededError(
+                session_key=state.config.session_key,
+                estimated_tokens=measured,
+                input_budget=self.input_budget(state.config),
+                source=source,
+            )
+
+        compaction.active_summary = summary
+        prepared = self.prepare_messages_for_model(
+            state.config,
+            [
+                *self._summary_transcript(compaction, summary),
+                {"role": "user", "content": SUMMARY_CONTINUATION_TEXT},
+                *delta_messages,
+            ],
+        )
+        # Responses-style state is append-only. Replacing H with a
+        # checkpoint requires a fresh request; a successful response may
+        # establish a new provider-owned state at the rewritten boundary.
+        state.conversation.replace_transcript(compaction.raw_messages)
+        state.usage = None
+        prepared = self.ensure_request_fits(
+            state.config,
+            prepared,
+            tool_definitions=tool_definitions,
+        )
+        compaction.summary_checkpoint = SessionSummaryCheckpoint(
+            summary=summary,
+            transcript_boundary=compaction.raw_accepted_boundary,
+        )
+        return prepared
+
+    async def prepare_request(
+        self,
+        state: ModelRequestState,
+        messages: list[dict[str, Any]],
+        *,
+        tool_definitions: list[dict[str, Any]] | None,
+        transcript: list[dict[str, Any]] | None = None,
+    ) -> tuple[list[dict[str, Any]], ProviderCallContext | None]:
+        """Prepare, compact or fit, and record the exact provider payload."""
+        prepared = self.prepare_messages_for_model(state.config, messages)
+        model_messages: list[dict[str, Any]] | None = prepared
+        supplemental_messages: list[dict[str, Any]] | None = None
+        request_context_tokens = None
+        if transcript is not None:
+            if tool_definitions is None:
+                model_messages = None
+                supplemental_messages = [prepared[-1]]
+            request_context_tokens = state.conversation.estimate_request_context_tokens(
+                transcript,
+                model_messages=model_messages,
+                supplemental_messages=supplemental_messages,
+                tool_definitions=tool_definitions,
+            )
+        usage_matches_messages = (
+            state.messages is not None
+            and prepared == state.messages
+            and tool_definitions == state.tool_definitions
+        )
+        request_was_fitted = False
+        compaction = state.compaction
+        if compaction is None:
+            prepared, request_was_fitted = self.fit_request(
+                state.config,
+                prepared,
+                state.usage,
+                usage_matches_messages=usage_matches_messages,
+                tool_definitions=tool_definitions,
+                request_context_tokens=request_context_tokens,
+            )
+        else:
+            pressure = self.request_pressure(
+                state.config,
+                prepared,
+                state.usage,
+                usage_matches_messages=usage_matches_messages,
+                tool_definitions=tool_definitions,
+                request_context_tokens=request_context_tokens,
+            )
+            if pressure is not None:
+                prepared = await self._compact_request_history(
+                    state,
+                    compaction,
+                    messages,
+                    pressure,
+                    tool_definitions=tool_definitions,
+                )
+                model_messages = prepared
+                supplemental_messages = None
+        provider_context = (
+            state.conversation.prepare_request(
+                transcript,
+                context_window_tokens=state.config.context_window_tokens,
+                model_messages=model_messages,
+                supplemental_messages=supplemental_messages,
+                resume_state=not request_was_fitted,
+            )
+            if transcript is not None
+            else state.conversation.independent_request_context(
+                context_window_tokens=state.config.context_window_tokens,
+            )
+        )
+        state.messages = deepcopy(prepared)
+        state.tool_definitions = deepcopy(tool_definitions)
+        return prepared, provider_context
+
     @staticmethod
     def input_budget(config: ContextGovernanceConfig) -> int:
         if not config.context_window_tokens:
@@ -424,14 +886,15 @@ class ContextGovernor:
         if budget <= 0:
             return messages
 
-        estimate, _ = estimate_prompt_tokens_chain(
-            config.provider,
-            config.model,
-            messages,
-            tool_definitions,
-        )
-        if not force and estimate <= budget:
-            return messages
+        if not force:
+            estimate, _ = estimate_prompt_tokens_chain(
+                config.provider,
+                config.model,
+                messages,
+                tool_definitions,
+            )
+            if estimate <= budget:
+                return messages
 
         system_messages = [dict(msg) for msg in messages if msg.get("role") == "system"]
         non_system = [dict(msg) for msg in messages if msg.get("role") != "system"]

nanobot/agent/loop.py

@@ -13,6 +13,7 @@ import weakref
 from collections.abc import Coroutine, Iterable, Mapping
 from contextlib import AbstractContextManager, ExitStack, nullcontext, suppress
 from dataclasses import dataclass, field
+from datetime import datetime
 from enum import Enum, auto
 from functools import partial
 from pathlib import Path
@@ -93,7 +94,11 @@ from nanobot.session.recovery import (
     restore_pending_interruption,
     restore_runtime_checkpoint,
 )
-from nanobot.session.summary import SessionSummary
+from nanobot.session.summary import (
+    SUMMARY_CONTINUATION_TEXT,
+    SessionSummary,
+    SessionSummaryCheckpoint,
+)
 from nanobot.triggers.local_turns import LocalTriggerTurnCoordinator
 from nanobot.utils.cancellation import task_is_cancelling
 from nanobot.utils.document import reference_non_image_attachments
@@ -161,6 +166,8 @@ class TurnContext:
 
     pending_queue: asyncio.Queue[InboundMessage] | None = None
     pending_summary: SessionSummary | None = None
+    summary_checkpoint: SessionSummaryCheckpoint | None = None
+    provider_compaction_applied: bool = False
 
     ephemeral: bool = False
     run_extra_hooks_for_ephemeral: bool = False
@@ -923,19 +930,6 @@ class AgentLoop:
             return
         remember_last_channel(session.metadata, msg.channel, msg.chat_id)
 
-    @staticmethod
-    def _replay_token_budget(runtime: LLMRuntime) -> int:
-        """Derive a token budget for session history replay from the context window."""
-        if runtime.context_window_tokens <= 0:
-            return 0
-        max_output = runtime.generation.max_tokens
-        try:
-            reserved_output = int(max_output)
-        except (TypeError, ValueError):
-            reserved_output = 4096
-        budget = runtime.context_window_tokens - max(1, reserved_output) - 1024
-        return budget if budget > 0 else max(128, runtime.context_window_tokens // 2)
-
     async def _run_agent_loop(
         self,
         transcript_input: TranscriptInput,
@@ -1186,6 +1180,26 @@ class AgentLoop:
                 provider_retry_mode=self.provider_retry_mode,
                 retry_wait_callback=on_retry_wait,
                 checkpoint_callback=_checkpoint,
+                consolidate_history=(
+                    partial(
+                        self.consolidator.summarize_transcript,
+                        runtime=runtime,
+                        session_key=session.key,
+                        tools=effective_tools.get_definitions(),
+                    )
+                    if session is not None and not ephemeral
+                    else None
+                ),
+                consolidate_provider_compaction=(
+                    partial(
+                        self.consolidator.summarize_provider_compaction,
+                        runtime=runtime,
+                        session_key=session.key,
+                        tools=effective_tools.get_definitions(),
+                    )
+                    if session is not None and not ephemeral
+                    else None
+                ),
                 injection_callback=_drain_pending,
                 terminal_injection_callback=_wait_for_pending,
                 # Sustained goals may legitimately exceed NANOBOT_LLM_TIMEOUT_S; idle stall
@@ -1886,12 +1900,6 @@ class AgentLoop:
         if ctx.on_runtime_admitted is not None:
             await ctx.on_runtime_admitted(runtime)
         if not ctx.ephemeral:
-            await self.consolidator.maybe_consolidate_by_tokens(
-                session,
-                runtime=runtime,
-            )
-            # Token consolidation may have committed a replacement checkpoint
-            # after the compact stage captured its summary for this request.
             ctx.session, ctx.pending_summary = self.auto_compact.prepare_session(
                 session,
                 ctx.session_key,
@@ -1899,11 +1907,7 @@ class AgentLoop:
             session = ctx.require_session()
         is_subagent = ctx.kind is TurnKind.SYSTEM and ctx.msg.sender_id == "subagent"
 
-        _hist_kwargs: dict[str, Any] = {
-            "max_tokens": self._replay_token_budget(runtime),
-            "extend_to_user": is_subagent,
-        }
-        ctx.history = session.get_history(**_hist_kwargs)
+        ctx.history = session.get_history(extend_to_user=is_subagent)
         stored_state = session.provider_state
         subagent_followup_persisted = False
         if is_subagent:
@@ -2021,6 +2025,8 @@ class AgentLoop:
             )
         ctx.final_content = result.final_content
         ctx.all_messages = result.messages
+        ctx.summary_checkpoint = result.summary_checkpoint
+        ctx.provider_compaction_applied = result.provider_compaction_applied
         ctx.stop_reason = result.stop_reason
         if (
             ctx.kind is TurnKind.USER
@@ -2034,7 +2040,6 @@ class AgentLoop:
             await turn_continuation.maybe_continue_turn(ctx)
 
     async def _persist_turn(self, ctx: TurnContext) -> None:
-        runtime = ctx.require_runtime()
         session = ctx.require_session()
         turn_continuation.prepare_save_boundary(ctx)
 
@@ -2060,15 +2065,18 @@ class AgentLoop:
         self._save_turn(
             session, ctx.all_messages, ctx.save_skip,
             turn_latency_ms=ctx.turn_latency_ms,
+            summary_checkpoint=ctx.summary_checkpoint,
+            input_persisted_early=ctx.input_persisted_early,
         )
+        if (
+            not ctx.ephemeral
+            and ctx.provider_compaction_applied
+            and ctx.summary_checkpoint is not None
+        ):
+            # The next request must rebuild from the portable checkpoint;
+            # the opaque continuation predates that transcript rewrite.
+            session.provider_state = None
         ctx.delivery.record_latency(ctx.turn_latency_ms)
-        if not ctx.ephemeral:
-            self.schedule_background(
-                self.consolidator.maybe_consolidate_by_tokens(
-                    session,
-                    runtime=runtime,
-                )
-            )
         self._clear_pending_user_turn(session)
         self._clear_runtime_checkpoint(session)
         self.sessions.save(session)
@@ -2142,6 +2150,55 @@ class AgentLoop:
 
         return filtered
 
+    @staticmethod
+    def _insert_summary_checkpoint(
+        session: Session,
+        checkpoint: SessionSummaryCheckpoint,
+        *,
+        insert_at: int | None = None,
+    ) -> None:
+        """Commit a replacement summary and its hidden transcript boundary."""
+        hint = {
+            "role": "user",
+            "content": SUMMARY_CONTINUATION_TEXT,
+            HIDDEN_HISTORY_META: True,
+            "timestamp": datetime.now().isoformat(),
+        }
+        if insert_at is None:
+            session.messages.append(hint)
+            checkpoint_session_index = len(session.messages) - 1
+        else:
+            session.messages.insert(insert_at, hint)
+            checkpoint_session_index = insert_at
+        session.metadata["_last_summary"] = {
+            "text": checkpoint.summary,
+            "last_active": session.updated_at.isoformat(),
+        }
+        session.last_archived = checkpoint_session_index
+
+    @staticmethod
+    def _validated_checkpoint_boundary(
+        checkpoint: SessionSummaryCheckpoint | None,
+        *,
+        skip: int,
+        message_count: int,
+        session_key: str,
+    ) -> int | None:
+        """Return a checkpoint boundary only when it belongs to this turn."""
+        if checkpoint is None:
+            return None
+        boundary = checkpoint.transcript_boundary
+        if skip - 1 <= boundary <= message_count:
+            return boundary
+        logger.warning(
+            "Ignoring invalid summary boundary {} outside [{}, {}] for {}",
+            boundary,
+            skip - 1,
+            message_count,
+            session_key,
+        )
+        return None
+
     def _save_turn(
         self,
         session: Session,
@@ -2149,10 +2206,10 @@ class AgentLoop:
         skip: int,
         *,
         turn_latency_ms: int | None = None,
+        summary_checkpoint: SessionSummaryCheckpoint | None = None,
+        input_persisted_early: bool = False,
     ) -> None:
-        """Save new-turn messages into session, truncating large tool results."""
-        from datetime import datetime
-
+        """Commit new-turn messages and an optional summary boundary."""
         declared_tool_call_ids = {
             str(tc["id"])
             for m in session.messages
@@ -2169,8 +2226,30 @@ class AgentLoop:
         }
         last_assistant_idx: int | None = None
         saved_followup_ids: set[str] = set()
-        for m in messages[skip:]:
-            entry = dict(m)
+        checkpoint_boundary = self._validated_checkpoint_boundary(
+            summary_checkpoint,
+            skip=skip,
+            message_count=len(messages),
+            session_key=session.key,
+        )
+
+        # The trigger input may already be the session tail while still being
+        # the first message after the replacement checkpoint.
+        if summary_checkpoint is not None and checkpoint_boundary == skip - 1:
+            insert_at = len(session.messages) - (1 if input_persisted_early else 0)
+            self._insert_summary_checkpoint(
+                session,
+                summary_checkpoint,
+                insert_at=insert_at,
+            )
+
+        for message_index, message in enumerate(messages[skip:], start=skip):
+            # Insert against the raw transcript index before filtering the
+            # message so persistence cleanup cannot shift the H/Δ boundary.
+            if summary_checkpoint is not None and checkpoint_boundary == message_index:
+                self._insert_summary_checkpoint(session, summary_checkpoint)
+
+            entry = dict(message)
             followup_id_value = cast(object, entry.pop(PENDING_FOLLOWUP_ID_KEY, None))
             followup_ids = (
                 [followup_id_value]
@@ -2249,6 +2328,8 @@ class AgentLoop:
                     for tc in (cast(dict[str, Any], tc_value),)
                     if tc.get("id")
                 )
+        if summary_checkpoint is not None and checkpoint_boundary == len(messages):
+            self._insert_summary_checkpoint(session, summary_checkpoint)
         if turn_latency_ms is not None and last_assistant_idx is not None:
             session.messages[last_assistant_idx]["latency_ms"] = int(turn_latency_ms)
         if saved_followup_ids:

nanobot/agent/memory.py

@@ -1,4 +1,4 @@
-"""Memory storage, transcript archiving, and legacy consolidation coordination."""
+"""Memory storage, transcript archiving, and session checkpoint consolidation."""
 
 # Tool schemas are installed by the ``@tool_parameters`` class decorator at
 # runtime; static analyzers cannot observe that it clears ``parameters`` from
@@ -21,6 +21,7 @@ from typing import TYPE_CHECKING, Any, Callable, Iterator, cast
 from loguru import logger
 
 from nanobot.llm_usage.context import llm_usage_source
+from nanobot.providers.base import ProviderCallContext, ProviderConversationState
 from nanobot.runtime_context import public_history_messages
 from nanobot.session.manager import (
     MIN_COMPACTED_REPLAY_MESSAGES,
@@ -740,7 +741,7 @@ class MemoryStore:
 
 
 # ---------------------------------------------------------------------------
-# Memory ingestion and legacy context-pressure coordination
+# Memory ingestion and context-pressure coordination
 # ---------------------------------------------------------------------------
 
 # Raw fallbacks use a tighter cap. Completed model summaries may scale with the
@@ -780,6 +781,20 @@ class MemoryArchiver:
     ) -> str:
         """Persist the failed chunk and return a bounded replacement checkpoint."""
         raw = self.store.raw_archive(messages, session_key=session_key)
+        return self._combine_raw_checkpoint(
+            raw,
+            previous_summary=previous_summary,
+            max_tokens=max_tokens,
+        )
+
+    @staticmethod
+    def _combine_raw_checkpoint(
+        raw: str,
+        *,
+        previous_summary: str | None,
+        max_tokens: int,
+    ) -> str:
+        """Return a bounded checkpoint that preserves prior and newly archived context."""
         token_limit = max(1, max_tokens)
         if not previous_summary:
             return truncate_text_to_tokens(raw, token_limit)
@@ -806,35 +821,94 @@ class MemoryArchiver:
 
     async def archive(
         self,
-        messages: list[dict[str, Any]],
+        source_messages: list[dict[str, Any]],
         *,
         runtime: LLMRuntime,
         session_key: str,
-        request_messages: list[dict[str, Any]],
+        history: list[dict[str, Any]],
         request_tools: list[dict[str, Any]],
         previous_summary: str | None = None,
+        input_token_budget: int | None = None,
+        fallback_max_tokens: int | None = None,
+        provider_state: ProviderConversationState | None = None,
     ) -> str | None:
-        """Execute a prepared archive request and persist its result."""
-        if not messages:
+        """Append the archive prompt to H and persist its summary."""
+        if not source_messages:
             return None
 
         def raw_fallback() -> str:
             return self._raw_checkpoint(
-                messages,
+                source_messages,
                 session_key=session_key,
                 previous_summary=previous_summary,
-                max_tokens=runtime.generation.max_tokens,
+                max_tokens=(
+                    fallback_max_tokens
+                    if fallback_max_tokens is not None
+                    else runtime.generation.max_tokens
+                ),
             )
 
+        prompt = render_template(
+            "agent/consolidator_archive.md",
+            strip=True,
+            archive_count=len(source_messages),
+        )
+        prompt_message = {"role": "user", "content": prompt}
+        provider_context = None
+        call_tools = request_tools
+        if provider_state is not None:
+            if not runtime.provider.can_resume_conversation_state(
+                provider_state,
+                runtime.model,
+            ):
+                return raw_fallback()
+            instruction_messages: list[dict[str, Any]] = []
+            for message in history:
+                if message.get("role") not in {"system", "developer"}:
+                    break
+                instruction_messages.append(dict(message))
+            request_messages = [*instruction_messages, prompt_message]
+            provider_context = ProviderCallContext(
+                conversation_state=provider_state.with_pending_messages([
+                    *provider_state.pending_messages,
+                    prompt_message,
+                ]),
+                context_window_tokens=runtime.context_window_tokens,
+                session_id=session_key,
+            )
+            call_tools = []
+        else:
+            request_messages = [
+                *[dict(message) for message in history],
+                prompt_message,
+            ]
+        if input_token_budget is not None and provider_context is None:
+            estimated, source = estimate_prompt_tokens_chain(
+                runtime.provider,
+                runtime.model,
+                request_messages,
+                call_tools,
+            )
+            if input_token_budget <= 0 or estimated > input_token_budget:
+                logger.debug(
+                    "Memory archive input does not fit for {}: {}/{} via {}; raw-dumping",
+                    session_key,
+                    estimated,
+                    input_token_budget,
+                    source,
+                )
+                return raw_fallback()
+
         try:
             with llm_usage_source("dream"):
                 response = await runtime.provider.chat_with_retry(
                     model=runtime.model,
                     messages=request_messages,
-                    tools=request_tools,
+                    tools=call_tools,
                     temperature=runtime.generation.temperature,
                     max_tokens=runtime.generation.max_tokens,
                     reasoning_effort=runtime.generation.reasoning_effort,
+                    provider_context=provider_context,
                 )
         except Exception:
             logger.warning("Memory archive provider call failed, raw-dumping to history")
@@ -879,20 +953,17 @@ class MemoryArchiver:
         )
         previous_summary = session_summary["text"] if session_summary else None
 
-        def raw_fallback() -> str:
+        if input_token_budget <= 0:
+            logger.debug(
+                "Memory archive has no safe input budget for {}; raw-dumping",
+                session.key,
+            )
             return self._raw_checkpoint(
                 messages,
                 session_key=session.key,
                 previous_summary=previous_summary,
                 max_tokens=runtime.generation.max_tokens,
             )
-
-        if input_token_budget <= 0:
-            logger.debug(
-                "Memory archive has no safe input budget for {}; raw-dumping",
-                session.key,
-            )
-            return raw_fallback()
         prefix = Session(
             key=session.key,
             messages=list(session.messages[:archive_end]),
@@ -908,47 +979,37 @@ class MemoryArchiver:
                 "Memory archive cannot replay the full chunk for {}; raw-dumping",
                 session.key,
             )
-            return raw_fallback()
-        prompt = render_template("agent/consolidator_archive.md", strip=True)
+            return self._raw_checkpoint(
+                messages,
+                session_key=session.key,
+                previous_summary=previous_summary,
+                max_tokens=runtime.generation.max_tokens,
+            )
         channel = session.key.split(":", 1)[0] if ":" in session.key else None
         workspace: Path | None = None
         if self._resolve_prompt_context is not None:
             channel, workspace = self._resolve_prompt_context(session)
-        request_messages = self._build_messages(
+        history_messages = self._build_messages(
             history=history,
-            current_message=prompt,
+            current_message=None,
             channel=channel,
             session_summary=session_summary,
             workspace=workspace,
         )
         tools = self._get_tool_definitions()
-        estimated, source = estimate_prompt_tokens_chain(
-            runtime.provider,
-            runtime.model,
-            request_messages,
-            tools,
-        )
-        if estimated > input_token_budget:
-            logger.debug(
-                "Memory archive prefix exceeds budget for {}; raw-dumping: {}/{} via {}",
-                session.key,
-                estimated,
-                input_token_budget,
-                source,
-            )
-            return raw_fallback()
         return await self.archive(
             messages,
             runtime=runtime,
             session_key=session.key,
-            request_messages=request_messages,
+            history=history_messages,
             request_tools=tools,
             previous_summary=previous_summary,
+            input_token_budget=input_token_budget,
         )
 
 
 class Consolidator:
-    """Legacy context-pressure coordinator backed by a MemoryArchiver."""
+    """Coordinate session Memory checkpoints through ``MemoryArchiver``."""
 
     _SAFETY_BUFFER = 1024  # extra headroom for tokenizer estimation drift
 
@@ -978,22 +1039,73 @@ class Consolidator:
         """Return the shared consolidation lock for one session."""
         return self._locks.setdefault(session_key, asyncio.Lock())
 
-    def pick_consolidation_boundary(
+    async def summarize_transcript(
         self,
-        session: Session,
-    ) -> int | None:
-        """Return the fixed user-led boundary before the recent replay tail."""
-        if not session.messages:
+        accepted_messages: list[dict[str, Any]],
+        previous_summary: str | None,
+        *,
+        runtime: LLMRuntime,
+        session_key: str,
+        tools: list[dict[str, Any]],
+        provider_state: ProviderConversationState | None = None,
+    ) -> str | None:
+        """Summarize the exact transcript prefix already accepted by the model."""
+        source_messages = [
+            dict(message)
+            for message in accepted_messages
+            if message.get("role") != "system"
+        ]
+        if not source_messages:
             return None
-        boundary = max(0, len(session.messages) - MIN_COMPACTED_REPLAY_MESSAGES)
-        while boundary > 0 and session.messages[boundary].get("role") != "user":
-            boundary -= 1
-        if (
-            boundary <= session.last_archived
-            or session.messages[boundary].get("role") != "user"
-        ):
+
+        max_output_tokens = max(0, runtime.generation.max_tokens)
+        input_token_budget = runtime.context_window_tokens - max_output_tokens
+        checkpoint_tokens = min(
+            max_output_tokens,
+            max(1, (input_token_budget - self._SAFETY_BUFFER) // 2),
+        )
+
+        summary = await self.archiver.archive(
+            source_messages,
+            runtime=runtime,
+            session_key=session_key,
+            history=accepted_messages,
+            request_tools=tools,
+            previous_summary=previous_summary,
+            input_token_budget=input_token_budget,
+            fallback_max_tokens=max(1, checkpoint_tokens),
+            provider_state=provider_state,
+        )
+        if summary == "(nothing)":
+            summary = self.archiver._raw_checkpoint(
+                source_messages,
+                session_key=session_key,
+                previous_summary=previous_summary,
+                max_tokens=max_output_tokens,
+            )
+        if summary is None:
             return None
-        return boundary
+        return truncate_text_to_tokens(summary, max(1, max_output_tokens))
+
+    async def summarize_provider_compaction(
+        self,
+        state: ProviderConversationState,
+        fallback_messages: list[dict[str, Any]],
+        previous_summary: str | None,
+        *,
+        runtime: LLMRuntime,
+        session_key: str,
+        tools: list[dict[str, Any]],
+    ) -> str | None:
+        """Prompt a native compacted state without replaying its raw history."""
+        return await self.summarize_transcript(
+            fallback_messages,
+            previous_summary,
+            runtime=runtime,
+            session_key=session_key,
+            tools=tools,
+            provider_state=state,
+        )
 
     @staticmethod
     def _full_replay_history(
@@ -1058,7 +1170,7 @@ class Consolidator:
         archive_end: int,
         runtime: LLMRuntime,
     ) -> str | None:
-        """Compatibility wrapper for the extracted MemoryArchiver."""
+        """Archive one captured session range through the shared Memory path."""
         return await self.archiver.archive_session(
             session,
             archive_end=archive_end,
@@ -1066,78 +1178,6 @@ class Consolidator:
             input_token_budget=self._input_token_budget(runtime),
         )
 
-    async def maybe_consolidate_by_tokens(
-        self,
-        session: Session,
-        *,
-        runtime: LLMRuntime,
-    ) -> None:
-        """Archive one fixed old prefix when the prompt exceeds the safe budget.
-
-        The budget reserves space for completion tokens and a safety buffer
-        so the LLM request never exceeds the context window.
-        """
-        lock = self.get_lock(session.key)
-        async with lock:
-            # Refresh session reference: AutoCompact may have replaced it.
-            fresh = self.sessions.get_or_create(session.key)
-            if fresh is not session:
-                session = fresh
-            if runtime.context_window_tokens <= 0:
-                return
-            if not session.messages:
-                return
-
-            budget = self._input_token_budget(runtime)
-            estimated, source = self.estimate_session_prompt_tokens(
-                session,
-                runtime=runtime,
-            )
-            if estimated <= 0:
-                return
-            if estimated < budget:
-                unarchived_count = len(session.messages) - session.last_archived
-                logger.debug(
-                    "Token consolidation idle {}: {}/{} via {}, msgs={}",
-                    session.key,
-                    estimated,
-                    runtime.context_window_tokens,
-                    source,
-                    unarchived_count,
-                )
-                return
-
-            end_idx = self.pick_consolidation_boundary(session)
-            if end_idx is None:
-                logger.debug(
-                    "Token consolidation: no safe fixed boundary for {}",
-                    session.key,
-                )
-                return
-
-            chunk = session.messages[session.last_archived:end_idx]
-            if not chunk:
-                return
-
-            logger.info(
-                "Token consolidation for {}: {}/{} via {}, chunk={} msgs",
-                session.key,
-                estimated,
-                runtime.context_window_tokens,
-                source,
-                len(chunk),
-            )
-            summary = await self.archive_session(
-                session,
-                archive_end=end_idx,
-                runtime=runtime,
-            )
-            if summary is None:
-                return
-            self._set_last_summary(session, summary)
-            session.last_archived = end_idx
-            self.sessions.save(session)
-
     async def compact_idle_session(
         self,
         session_key: str,

nanobot/agent/runner.py

@@ -16,8 +16,13 @@ from loguru import logger
 
 from nanobot.agent.context import TranscriptInput
 from nanobot.agent.context_governance import (
+    ContextCompactionState,
     ContextGovernanceConfig,
     ContextGovernor,
+    HistoryConsolidator,
+    ModelRequestState,
+    ProviderCompactionConsolidator,
+    TranscriptBuilder,
 )
 from nanobot.agent.hook import AgentHook, AgentHookContext, AgentRunHookContext
 from nanobot.agent.tools.execution import execute_tool_calls
@@ -32,20 +37,10 @@ from nanobot.providers.base import (
     LLMProvider,
     LLMResponse,
     LLMUsage,
-    ProviderCallContext,
     ProviderConversationState,
 )
-from nanobot.providers.conversation_state import (
-    ProviderConversationStateController,
-    allows_conversation_message_merge,
-)
-from nanobot.runtime_context import (
-    RUNTIME_CONTEXT_MESSAGE_META,
-    detach_runtime_context,
-    reattach_runtime_context,
-)
-from nanobot.session.history_visibility import is_hidden_history_message
-from nanobot.session.recovery import PENDING_FOLLOWUP_ID_KEY
+from nanobot.providers.conversation_state import ProviderConversationStateController
+from nanobot.session.summary import SessionSummaryCheckpoint
 from nanobot.utils.helpers import (
     build_assistant_message,
     estimate_message_tokens,
@@ -67,7 +62,6 @@ ContinuationCallback = Callable[[], str | None]
 RetryWaitCallback = Callable[[str], Awaitable[None]]
 CheckpointCallback = Callable[[dict[str, Any]], Awaitable[None]]
 InjectionCallback = Callable[..., Awaitable[Iterable[Any] | None]]
-TranscriptBuilder = Callable[[TranscriptInput], list[dict[str, Any]]]
 
 _DEFAULT_ERROR_MESSAGE = "Sorry, I encountered an error calling the AI model."
 _ARREARAGE_ERROR_MESSAGE = (
@@ -113,6 +107,8 @@ class AgentRunSpec:
     provider_retry_mode: str = "standard"
     retry_wait_callback: RetryWaitCallback | None = None
     checkpoint_callback: CheckpointCallback | None = None
+    consolidate_history: HistoryConsolidator | None = None
+    consolidate_provider_compaction: ProviderCompactionConsolidator | None = None
     injection_callback: InjectionCallback | None = None
     terminal_injection_callback: InjectionCallback | None = None
     llm_timeout_s: float | None = None
@@ -137,17 +133,8 @@ class AgentRunResult:
     # Terminal tail to emit when the preceding final-content prefix was already streamed.
     pending_stream_content: str | None = None
     provider_state: ProviderConversationState | None = field(default=None, repr=False)
-
-
-@dataclass(slots=True)
-class _ModelRequestState:
-    """Per-run state used to govern the next provider request."""
-
-    config: ContextGovernanceConfig
-    conversation: ProviderConversationStateController
-    usage: LLMUsage | None = None
-    messages: list[dict[str, Any]] | None = None
-    tool_definitions: list[dict[str, Any]] | None = None
+    summary_checkpoint: SessionSummaryCheckpoint | None = field(default=None, repr=False)
+    provider_compaction_applied: bool = field(default=False, repr=False)
 
 
 class AgentRunner:
@@ -157,118 +144,12 @@ class AgentRunner:
         self.context_governor = ContextGovernor()
 
     @staticmethod
-    def _merge_message_content(left: Any, right: Any) -> str | list[dict[str, Any]]:
-        if isinstance(left, str) and isinstance(right, str):
-            return f"{left}\n\n{right}" if left else right
-
-        def _to_blocks(value: Any) -> list[dict[str, Any]]:
-            if isinstance(value, list):
-                return [
-                    cast(dict[str, Any], item)
-                    if isinstance(item, dict)
-                    else {"type": "text", "text": str(item)}
-                    for item in cast(list[Any], value)
-                ]
-            if value is None:
-                return []
-            return [{"type": "text", "text": str(value)}]
-
-        return _to_blocks(left) + _to_blocks(right)
-
-    @classmethod
     def _append_injected_messages(
-        cls,
         messages: list[dict[str, Any]],
         injections: list[dict[str, Any]],
     ) -> None:
-        """Append injected user messages while preserving role alternation."""
-        for injection in injections:
-            if (
-                messages
-                and injection.get("role") == "user"
-                and messages[-1].get("role") == "user"
-                and not is_hidden_history_message(injection)
-                and not is_hidden_history_message(messages[-1])
-                and allows_conversation_message_merge(messages[-1])
-            ):
-                merged = dict(messages[-1])
-                left_meta = merged.get("_meta")
-                right_meta = injection.get("_meta")
-                left_meta_dict = cast(dict[str, Any], left_meta) if isinstance(left_meta, dict) else None
-                right_meta_dict = (
-                    cast(dict[str, Any], right_meta) if isinstance(right_meta, dict) else None
-                )
-                left_marker = (
-                    left_meta_dict.get(RUNTIME_CONTEXT_MESSAGE_META)
-                    if left_meta_dict is not None
-                    else None
-                )
-                right_marker = (
-                    right_meta_dict.get(RUNTIME_CONTEXT_MESSAGE_META)
-                    if right_meta_dict is not None
-                    else None
-                )
-                left_marker_dict = (
-                    cast(dict[str, Any], left_marker) if isinstance(left_marker, dict) else None
-                )
-                right_marker_dict = (
-                    cast(dict[str, Any], right_marker) if isinstance(right_marker, dict) else None
-                )
-                empty_sources: list[str] = []
-                empty_blocks: list[dict[str, Any]] = []
-                detached_left = (
-                    detach_runtime_context(merged.get("content"), left_marker_dict)
-                    if left_marker_dict is not None
-                    else (merged.get("content"), empty_sources, empty_blocks)
-                )
-                detached_right = (
-                    detach_runtime_context(injection.get("content"), right_marker_dict)
-                    if right_marker_dict is not None
-                    else (injection.get("content"), empty_sources, empty_blocks)
-                )
-                if detached_left is not None and detached_right is not None:
-                    left_content, left_sources, left_blocks = detached_left
-                    right_content, right_sources, right_blocks = detached_right
-                    merged_content = cls._merge_message_content(left_content, right_content)
-                    context_blocks = [*left_blocks, *right_blocks]
-                    if context_blocks:
-                        merged_content, marker = reattach_runtime_context(
-                            merged_content,
-                            [*left_sources, *right_sources],
-                            context_blocks,
-                        )
-                        internal_meta = dict(left_meta_dict) if left_meta_dict is not None else {}
-                        if right_meta_dict is not None:
-                            for key, value in right_meta_dict.items():
-                                internal_meta.setdefault(key, value)
-                        internal_meta[RUNTIME_CONTEXT_MESSAGE_META] = marker
-                        merged["_meta"] = internal_meta
-                    merged["content"] = merged_content
-                else:
-                    merged["content"] = cls._merge_message_content(
-                        merged.get("content"),
-                        injection.get("content"),
-                    )
-                followup_id = injection.get(PENDING_FOLLOWUP_ID_KEY)
-                if isinstance(followup_id, str) and followup_id:
-                    existing = cast(object, merged.get(PENDING_FOLLOWUP_ID_KEY))
-                    followup_ids = (
-                        [existing]
-                        if isinstance(existing, str)
-                        else [
-                            item
-                            for item in cast(list[object], existing)
-                            if isinstance(item, str)
-                        ]
-                        if isinstance(existing, list)
-                        else []
-                    )
-                    if followup_id not in followup_ids:
-                        followup_ids.append(followup_id)
-                    merged[PENDING_FOLLOWUP_ID_KEY] = followup_ids
-                messages[-1] = merged
-                continue
-            messages.append(injection)
+        """Append injected messages without rewriting the raw transcript."""
+        messages.extend(injections)
 
     async def _try_drain_injections(
         self,
@@ -425,7 +306,7 @@ class AgentRunner:
 
     async def run(self, spec: AgentRunSpec) -> AgentRunResult:
         hook = spec.hook or AgentHook()
-        messages = self._initial_transcript(spec)
+        messages, compaction = self._initial_transcript_and_compaction(spec)
         context = AgentRunHookContext(messages=deepcopy(messages))
         llm_usage_source_token = bind_llm_usage_source(
             spec.llm_usage_source or source_from_session_key(spec.session_key)
@@ -433,7 +314,7 @@ class AgentRunner:
 
         try:
             await hook.before_run(context)
-            result = await self._run_core(spec, hook, messages)
+            result = await self._run_core(spec, hook, messages, compaction)
         except asyncio.CancelledError as exc:
             context.messages = deepcopy(messages)
             context.stop_reason = "cancelled"
@@ -478,23 +359,35 @@ class AgentRunner:
                 reset_llm_usage_source(llm_usage_source_token)
 
     @staticmethod
-    def _initial_transcript(spec: AgentRunSpec) -> list[dict[str, Any]]:
-        """Resolve exactly one supported source for the initial model transcript."""
-        if spec.transcript_input is not None:
+    def _initial_transcript_and_compaction(
+        spec: AgentRunSpec,
+    ) -> tuple[list[dict[str, Any]], ContextCompactionState | None]:
+        """Build the initial transcript and its optional compaction state."""
+        transcript_input = spec.transcript_input
+        if transcript_input is not None:
             if spec.initial_messages is not None:
                 raise ValueError("provide either transcript_input or initial_messages, not both")
-            if spec.transcript_builder is None:
+            transcript_builder = spec.transcript_builder
+            if transcript_builder is None:
                 raise ValueError("transcript_builder is required with transcript_input")
-            return list(spec.transcript_builder(spec.transcript_input))
+            return ContextCompactionState.from_transcript(
+                transcript_input,
+                transcript_builder,
+                spec.consolidate_history,
+                spec.consolidate_provider_compaction,
+            )
         if spec.initial_messages is None:
             raise ValueError("initial_messages is required without transcript_input")
-        return list(spec.initial_messages)
+        if spec.consolidate_history is not None:
+            raise ValueError("consolidate_history requires transcript_input")
+        return list(spec.initial_messages), None
 
     async def _run_core(
         self,
         spec: AgentRunSpec,
         hook: AgentHook,
         messages: list[dict[str, Any]],
+        compaction: ContextCompactionState | None,
     ) -> AgentRunResult:
         final_content: str | None = None
         tools_used: list[str] = []
@@ -530,9 +423,10 @@ class AgentRunner:
             context_block_limit=spec.context_block_limit,
             max_tokens=spec.runtime.generation.max_tokens,
         )
-        request_state = _ModelRequestState(
+        request_state = ModelRequestState(
             config=governance_config,
             conversation=conversation_state,
+            compaction=compaction,
         )
 
         for iteration in range(spec.max_iterations):
@@ -542,9 +436,15 @@ class AgentRunner:
                 session_key=spec.session_key,
             )
             await hook.before_iteration(context)
+            request_message_count = len(messages)
+            request_messages = (
+                request_state.compaction.request_messages(messages)
+                if request_state.compaction is not None
+                else messages
+            )
             response = await self._request_model(
                 spec,
-                messages,
+                request_messages,
                 hook,
                 context,
                 request_state=request_state,
@@ -553,6 +453,11 @@ class AgentRunner:
             assert request_state.messages is not None
             messages_for_model = request_state.messages
             conversation_state.observe_response(response, messages)
+            if request_state.compaction is not None:
+                request_state.compaction.accept_request(
+                    messages_for_model,
+                    raw_boundary=request_message_count,
+                )
             context.response = response
             context.tool_calls = list(response.tool_calls)
 
@@ -634,7 +539,7 @@ class AgentRunner:
                     messages.append(tool_message)
                     completed_tool_results.append(tool_message)
                 checkpoint_model_messages = (
-                    self.context_governor.prepare_for_model(
+                    self.context_governor.prepare_messages_for_model(
                         governance_config,
                         messages,
                     )
@@ -920,6 +825,12 @@ class AgentRunner:
             had_injections=had_injections,
             pending_stream_content=pending_stream_content,
             provider_state=conversation_state.finish(messages),
+            summary_checkpoint=(
+                request_state.compaction.summary_checkpoint
+                if request_state.compaction is not None
+                else None
+            ),
+            provider_compaction_applied=request_state.provider_compaction_applied,
         )
 
     def _build_request_kwargs(
@@ -942,60 +853,6 @@ class AgentRunner:
         kwargs["reasoning_effort"] = generation.reasoning_effort
         return kwargs
 
-    def _prepare_model_request(
-        self,
-        state: _ModelRequestState,
-        messages: list[dict[str, Any]],
-        *,
-        tool_definitions: list[dict[str, Any]] | None,
-        transcript: list[dict[str, Any]] | None = None,
-    ) -> tuple[list[dict[str, Any]], ProviderCallContext | None]:
-        """Prepare, fit, and record the exact payload sent to a provider."""
-        prepared = self.context_governor.prepare_for_model(state.config, messages)
-        supplemental_messages = (
-            [prepared[-1]] if transcript is not None and tool_definitions is None else None
-        )
-        model_messages = None if supplemental_messages is not None else prepared
-        request_context_tokens = (
-            state.conversation.estimate_request_context_tokens(
-                transcript,
-                model_messages=model_messages,
-                supplemental_messages=supplemental_messages,
-                tool_definitions=tool_definitions,
-            )
-            if transcript is not None
-            else None
-        )
-        usage_matches_messages = (
-            state.messages is not None
-            and prepared == state.messages
-            and tool_definitions == state.tool_definitions
-        )
-        prepared, fitted = self.context_governor.fit_request(
-            state.config,
-            prepared,
-            state.usage,
-            usage_matches_messages=usage_matches_messages,
-            tool_definitions=tool_definitions,
-            request_context_tokens=request_context_tokens,
-        )
-        provider_context = (
-            state.conversation.prepare_request(
-                transcript,
-                context_window_tokens=state.config.context_window_tokens,
-                model_messages=model_messages,
-                supplemental_messages=supplemental_messages,
-                resume_state=not fitted,
-            )
-            if transcript is not None
-            else state.conversation.independent_request_context(
-                context_window_tokens=state.config.context_window_tokens,
-            )
-        )
-        state.messages = deepcopy(prepared)
-        state.tool_definitions = deepcopy(tool_definitions)
-        return prepared, provider_context
-
     async def _request_model(
         self,
         spec: AgentRunSpec,
@@ -1003,13 +860,13 @@ class AgentRunner:
         hook: AgentHook,
         context: AgentHookContext,
         *,
-        request_state: _ModelRequestState,
+        request_state: ModelRequestState,
         malformed_retry: bool = False,
         transcript: list[dict[str, Any]] | None,
     ) -> LLMResponse:
         timeout_s = self._resolve_llm_timeout_s(spec)
         tool_definitions = spec.tools.get_definitions()
-        messages, provider_context = self._prepare_model_request(
+        messages, provider_context = await self.context_governor.prepare_request(
             request_state,
             messages,
             tool_definitions=tool_definitions,
@@ -1169,6 +1026,12 @@ class AgentRunner:
             response.ttft_ms = max(0, round((first_output_at - request_started_at) * 1000))
         if generation_elapsed_s > 0:
             response.generation_ms = max(1, round(generation_elapsed_s * 1000))
+        await self.context_governor.summarize_provider_compaction(
+            request_state,
+            response,
+            current_request_boundary=(len(transcript) if transcript is not None else None),
+        )
+        request_state.provider_compaction_applied |= response.provider_compaction_applied
         # chat_stream_with_retry may recover internally, so only fail unfinished
         # hosted calls after the provider returns its final error response.
         if response.finish_reason == "error":
@@ -1283,7 +1146,7 @@ class AgentRunner:
         spec: AgentRunSpec,
         messages: list[dict[str, Any]],
         *,
-        request_state: _ModelRequestState,
+        request_state: ModelRequestState,
         transcript: list[dict[str, Any]],
     ) -> LLMResponse:
         retry_messages = self._finalization_retry_messages(messages)
@@ -1313,14 +1176,21 @@ class AgentRunner:
         messages: list[dict[str, Any]],
         usage: LLMUsage | None,
         *,
-        request_state: _ModelRequestState,
+        request_state: ModelRequestState,
     ) -> tuple[str | None, LLMUsage | None]:
-        retry_messages = self._budget_exhausted_finalization_messages(messages)
+        compaction = request_state.compaction
+        request_messages = (
+            compaction.request_messages(messages)
+            if compaction is not None
+            else messages
+        )
+        retry_messages = self._budget_exhausted_finalization_messages(request_messages)
         try:
             response = await self._request_no_tools(
                 spec,
                 retry_messages,
                 request_state=request_state,
+                transcript=messages if compaction is not None else None,
             )
         except Exception:
             logger.exception(
@@ -1358,10 +1228,10 @@ class AgentRunner:
         spec: AgentRunSpec,
         messages: list[dict[str, Any]],
         *,
-        request_state: _ModelRequestState,
+        request_state: ModelRequestState,
         transcript: list[dict[str, Any]] | None = None,
     ) -> LLMResponse:
-        messages, provider_context = self._prepare_model_request(
+        messages, provider_context = await self.context_governor.prepare_request(
             request_state,
             messages,
             tool_definitions=None,
@@ -1389,6 +1259,12 @@ class AgentRunner:
                 finish_reason="error",
                 error_kind="timeout",
             )
+        await self.context_governor.summarize_provider_compaction(
+            request_state,
+            response,
+            current_request_boundary=(len(transcript) if transcript is not None else None),
+        )
+        request_state.provider_compaction_applied |= response.provider_compaction_applied
         return response
 
     @staticmethod
@@ -1453,7 +1329,7 @@ class AgentRunner:
     def _record_request_usage(
         self,
         spec: AgentRunSpec,
-        state: _ModelRequestState,
+        state: ModelRequestState,
         response: LLMResponse,
     ) -> LLMUsage | None:
         assert state.messages is not None

nanobot/cli/gateway_runtime.py

@@ -662,11 +662,6 @@ def _run_gateway(
                 if isinstance(message_tool, MessageTool) and suppress_token is not None:
                     message_tool.reset_suppress_delivery(suppress_token)
 
-            # Keep a small tail of heartbeat history so the loop stays bounded.
-            session = agent.sessions.get_or_create("heartbeat")
-            session.retain_recent_legal_suffix(hb_cfg.keep_recent_messages)
-            agent.sessions.save(session)
-
             if not resp or not resp.content:
                 return
 

nanobot/config/schema.py

@@ -329,7 +329,6 @@ class HeartbeatConfig(Base):
 
     enabled: bool = True
     interval_s: int = 30 * 60  # 30 minutes
-    keep_recent_messages: int = 8
 
 
 class ApiConfig(Base):

nanobot/providers/azure_openai_provider.py

@@ -34,6 +34,7 @@ from nanobot.providers.base import (
 )
 from nanobot.providers.openai_responses import (
     ResponsesStreamCapture,
+    build_responses_compaction_state,
     build_responses_state,
     consume_sdk_stream,
     convert_tools,
@@ -410,6 +411,16 @@ class AzureOpenAIProvider(LLMProvider):
                     output_items=capture.output_items,
                     usage=usage,
                 )
+                result.provider_compaction_state = build_responses_compaction_state(
+                    provider=self._responses_state_provider(),
+                    model=str(body["model"]),
+                    output_items=capture.output_items,
+                )
+                result.provider_compaction_applied = (
+                    result.provider_compaction_state is not None
+                )
+                if result.provider_compaction_applied:
+                    result.provider_compaction_scope = "current_request"
             return result
         except Exception as e:
             return self._handle_error(e)

nanobot/providers/base.py

@@ -31,6 +31,7 @@ RETRY_AFTER_BUFFER = 1
 
 RetryEventCallback = Callable[[str], Awaitable[None]]
 LLMCallObserver = Callable[["LLMCallRecord"], None]
+ProviderCompactionScope = Literal["prior_context", "current_request"]
 
 
 def resolve_stream_idle_timeout_s(
@@ -563,6 +564,22 @@ class LLMResponse:
     reasoning_content: str | None = None  # Kimi, DeepSeek-R1, MiMo etc.
     thinking_blocks: list[dict[str, Any]] | None = None  # Anthropic extended thinking
     provider_state: ProviderConversationState | None = field(default=None, repr=False)
+    # True only when this response installed a new provider-native compaction
+    # boundary. Replaying an older compaction item does not set this flag.
+    provider_compaction_applied: bool = field(default=False, repr=False)
+    # State immediately after native compaction, before the normal response
+    # continues. An archive prompt can resume this state without replaying H.
+    provider_compaction_state: ProviderConversationState | None = field(
+        default=None,
+        repr=False,
+    )
+    # Which model input the native compaction state replaces. Providers that
+    # compact before attaching the current request delta report
+    # ``prior_context``; in-request compaction reports ``current_request``.
+    provider_compaction_scope: ProviderCompactionScope | None = field(
+        default=None,
+        repr=False,
+    )
     # Routing wrappers may preserve or discard an incoming provider-owned
     # continuation independently of the final fallback error's retry policy.
     preserve_provider_state_on_error: bool | None = field(default=None, repr=False)

nanobot/providers/conversation_state.py

@@ -101,6 +101,12 @@ class ProviderConversationStateController:
         )
         return context_tokens + max(0, delta_tokens)
 
+    def replace_transcript(self, messages: list[dict[str, Any]]) -> None:
+        """Discard append-only provider state after a transcript rewrite."""
+        self._state = None
+        self._boundary = len(messages)
+        self._request_messages = []
+
     def prepare_request(
         self,
         messages: list[dict[str, Any]],

nanobot/providers/openai_codex_provider.py

@@ -31,6 +31,7 @@ from nanobot.providers.oauth_model_catalog import (
 )
 from nanobot.providers.openai_responses import (
     ResponsesStreamCapture,
+    build_responses_compaction_state,
     build_responses_state,
     consume_sse_with_reasoning,
     convert_tools,
@@ -137,6 +138,8 @@ class OpenAICodexProvider(LLMProvider):
             body.update(self._extra_body)
 
         stage = "oauth_token"
+        native_compaction_applied = False
+        native_compaction_state: ProviderConversationState | None = None
         try:
             token = await asyncio.to_thread(get_codex_token, proxy=self.proxy)
             headers = _build_headers(cast(str, token.account_id), token.access)
@@ -187,9 +190,11 @@ class OpenAICodexProvider(LLMProvider):
                 and responses_state_context_tokens(sanitized_state) >= compact_threshold
             ):
                 stage = "codex_compaction"
+                history_items = responses_state_items(sanitized_state) or []
+                delta_items = input_items[len(history_items):]
                 compact_body = {
                     **body,
-                    "input": [*input_items, {"type": "compaction_trigger"}],
+                    "input": [*history_items, {"type": "compaction_trigger"}],
                 }
                 try:
                     compact_result = await _send(compact_body, emit_deltas=False)
@@ -205,9 +210,16 @@ class OpenAICodexProvider(LLMProvider):
                     }:
                         raise RuntimeError("Codex compaction returned no compaction item")
                     body["input"] = [
-                        *_retained_compaction_messages(input_items),
+                        *_retained_compaction_messages(history_items),
                         *compact_items,
+                        *delta_items,
                     ]
+                    native_compaction_state = build_responses_compaction_state(
+                        provider=self._responses_state_provider(),
+                        model=_strip_model_prefix(model),
+                        output_items=compact_items,
+                    )
+                    native_compaction_applied = True
                 except Exception as compact_error:
                     if is_compaction_compatibility_error(compact_error):
                         self._native_compaction_available = False
@@ -220,7 +232,14 @@ class OpenAICodexProvider(LLMProvider):
                     )
 
             stage = "codex_request"
-            return await _send(body, emit_deltas=True)
+            result = await _send(body, emit_deltas=True)
+            result.provider_compaction_applied = (
+                result.provider_compaction_applied or native_compaction_applied
+            )
+            if native_compaction_state is not None:
+                result.provider_compaction_state = native_compaction_state
+                result.provider_compaction_scope = "prior_context"
+            return result
         except Exception as e:
             response = _codex_error_response(e)
             exc_type = "CodexHTTPError" if isinstance(e, _CodexHTTPError) else type(e).__name__

nanobot/providers/openai_compat_provider.py

@@ -36,6 +36,7 @@ from nanobot.providers.base import (
 )
 from nanobot.providers.openai_responses import (
     ResponsesStreamCapture,
+    build_responses_compaction_state,
     build_responses_state,
     consume_sdk_stream,
     convert_tools,
@@ -2049,6 +2050,18 @@ class OpenAICompatProvider(LLMProvider):
                             output_items=capture.output_items,
                             usage=usage,
                         )
+                        result.provider_compaction_state = (
+                            build_responses_compaction_state(
+                                provider=self._responses_state_provider(),
+                                model=str(body["model"]),
+                                output_items=capture.output_items,
+                            )
+                        )
+                        result.provider_compaction_applied = (
+                            result.provider_compaction_state is not None
+                        )
+                        if result.provider_compaction_applied:
+                            result.provider_compaction_scope = "current_request"
                     return result
                 except Exception as responses_error:
                     if self._spec and self._spec.name == "github_copilot":

nanobot/providers/openai_responses/__init__.py

@@ -18,6 +18,7 @@ from nanobot.providers.openai_responses.parsing import (
     parse_response_output,
 )
 from nanobot.providers.openai_responses.state import (
+    build_responses_compaction_state,
     build_responses_state,
     is_compaction_compatibility_error,
     prepare_responses_input,
@@ -40,6 +41,7 @@ __all__ = [
     "is_replayable_finish_reason",
     "map_finish_reason",
     "parse_response_output",
+    "build_responses_compaction_state",
     "build_responses_state",
     "is_compaction_compatibility_error",
     "prepare_responses_input",

nanobot/providers/openai_responses/parsing.py

@@ -11,7 +11,10 @@ import httpx
 from loguru import logger
 
 from nanobot.providers.base import LLMResponse, LLMUsage, ToolCallRequest, parse_tool_arguments
-from nanobot.providers.openai_responses.state import build_responses_state
+from nanobot.providers.openai_responses.state import (
+    build_responses_compaction_state,
+    build_responses_state,
+)
 
 FINISH_REASON_MAP = {
     "completed": "stop",
@@ -655,6 +658,14 @@ def parse_response_output(
             output_items=output,
             usage=usage,
         )
+        result.provider_compaction_state = build_responses_compaction_state(
+            provider=state_provider,
+            model=state_model,
+            output_items=output,
+        )
+        result.provider_compaction_applied = result.provider_compaction_state is not None
+        if result.provider_compaction_applied:
+            result.provider_compaction_scope = "current_request"
     return result
 
 

nanobot/providers/openai_responses/state.py

@@ -108,6 +108,28 @@ def build_responses_state(
     )
 
 
+def build_responses_compaction_state(
+    *,
+    provider: str,
+    model: str,
+    output_items: list[dict[str, Any]],
+) -> ProviderConversationState | None:
+    """Return the state at the latest native compaction output boundary."""
+    latest = None
+    for index, item in enumerate(output_items):
+        if item.get("type") in _COMPACTION_ITEM_TYPES:
+            latest = index
+    if latest is None:
+        return None
+    return ProviderConversationState(
+        kind=RESPONSES_STATE_KIND,
+        provider=provider,
+        model=model,
+        version=RESPONSES_STATE_VERSION,
+        payload={_ITEMS_KEY: [deepcopy(output_items[latest])]},
+    )
+
+
 def responses_state_items(
     state: ProviderConversationState,
 ) -> list[dict[str, Any]] | None:

nanobot/sdk/clients.py

@@ -209,11 +209,11 @@ class RuntimeClient:
         return self._loop.runtime_events.subscribe(handler, SessionTurnPersisted)
 
     async def compact_session(self, session_key: str) -> SessionSnapshot:
-        """Run token consolidation for one session."""
+        """Archive one session through the shared idle-compaction path."""
         session = self._loop.sessions.get_or_create(session_key)
         runtime = self._loop.runtime_for_session(session)
-        await self._loop.consolidator.maybe_consolidate_by_tokens(
-            session,
+        await self._loop.consolidator.compact_idle_session(
+            session_key,
             runtime=runtime,
         )
         return snapshot_from_session(self._loop.sessions.get_or_create(session_key))

nanobot/session/manager.py

@@ -27,7 +27,9 @@ from nanobot.runtime_context import (
     RUNTIME_CONTEXT_HISTORY_META,
     public_history_message,
 )
+from nanobot.session.history_visibility import is_hidden_history_message
 from nanobot.session.model_selection import SESSION_MODEL_PRESET_METADATA_KEY
+from nanobot.session.summary import SUMMARY_CONTINUATION_TEXT
 from nanobot.utils.helpers import (
     content_with_media_breadcrumbs,
     ensure_dir,
@@ -262,12 +264,6 @@ def _metadata_title(metadata: object) -> str:
     return strip_think(title)
 
 
-@dataclass
-class RetentionResult:
-    dropped: list[dict[str, Any]]
-    already_consolidated_count: int
-
-
 @dataclass(frozen=True)
 class SessionPolicy:
     """Runtime rules that do not belong in durable session data."""
@@ -286,9 +282,7 @@ class Session:
     created_at: datetime = field(default_factory=datetime.now)
     updated_at: datetime = field(default_factory=datetime.now)
     metadata: dict[str, Any] = field(default_factory=dict)
-    # Legacy storage name for the Memory ingestion watermark. New code should
-    # use ``last_archived`` so this progress is not confused with model-context
-    # compaction. Keep the field while persisted sessions and SDK callers migrate.
+    # Keep the legacy storage name while persisted sessions and SDK callers migrate.
     last_consolidated: int = 0
     provider_state: ProviderConversationState | None = field(default=None, repr=False)
     policy: SessionPolicy = field(default_factory=SessionPolicy, repr=False, compare=False)
@@ -309,7 +303,7 @@ class Session:
 
     @property
     def last_archived(self) -> int:
-        """Number of transcript messages already written to the Memory journal."""
+        """End of the latest committed Memory checkpoint."""
         return self.last_consolidated
 
     @last_archived.setter
@@ -337,14 +331,17 @@ class Session:
     ) -> list[dict[str, Any]]:
         """Return recent replayable messages for LLM input.
 
-        A positive ``max_messages`` applies an explicit caller-owned count
-        limit. The normal model path relies on ``max_tokens`` instead.
+        A committed in-turn checkpoint replaces its old prefix with the stored
+        summary and resumes replay at a hidden continuation marker. A positive
+        ``max_messages`` applies an additional caller-owned count limit.
         """
         replay_start = self.last_archived
-        if replay_start:
-            # ``last_archived`` is archive progress, not a replay boundary.
-            # Keep a small raw suffix for continuity, extending back to the user
-            # that started an assistant/tool sequence when necessary.
+        resumes_from_checkpoint = (
+            replay_start < len(self.messages)
+            and is_hidden_history_message(self.messages[replay_start])
+            and self.messages[replay_start].get("content") == SUMMARY_CONTINUATION_TEXT
+        )
+        if replay_start and not resumes_from_checkpoint:
             recent_start = recent_message_start_index(
                 self.messages,
                 MIN_COMPACTED_REPLAY_MESSAGES,
@@ -485,110 +482,6 @@ class Session:
         self.updated_at = datetime.now()
         self.metadata.pop("_last_summary", None)
 
-    def retain_recent_legal_suffix(
-        self,
-        max_messages: int,
-        *,
-        extend_to_user: bool = False,
-    ) -> RetentionResult:
-        """Keep a legal recent suffix, optionally extending it back to a user turn.
-
-        Returns a RetentionResult with dropped messages and how many of those
-        were in the already-consolidated prefix. This method mutates
-        self.messages and self.last_archived in place.
-        """
-        if max_messages <= 0:
-            dropped = list(self.messages)
-            lc = self.last_archived
-            self.clear()
-            return RetentionResult(
-                dropped=dropped,
-                already_consolidated_count=min(lc, len(dropped)),
-            )
-        if len(self.messages) <= max_messages:
-            return RetentionResult(
-                dropped=[],
-                already_consolidated_count=0,
-            )
-
-        original = list(self.messages)
-        before_lc = self.last_archived
-
-        start_idx = max(0, len(self.messages) - max_messages)
-        if extend_to_user:
-            recovered_user = next(
-                (i for i in range(start_idx, -1, -1) if self.messages[i].get("role") == "user"),
-                None,
-            )
-            if recovered_user is not None:
-                start_idx = recovered_user
-                if start_idx > 0 and self.messages[start_idx - 1].get("_channel_delivery"):
-                    start_idx -= 1
-
-        retained = self.messages[start_idx:]
-
-        # Prefer starting at a user turn (or its preceding _channel_delivery) when one exists within the retained window.
-        first_user = next((i for i, m in enumerate(retained) if m.get("role") == "user"), None)
-        if first_user is not None:
-            if first_user > 0 and retained[first_user - 1].get("_channel_delivery"):
-                retained = retained[first_user - 1:]
-            else:
-                retained = retained[first_user:]
-        elif not extend_to_user:
-            # If the hard-capped tail is assistant/tool-only, anchor to the
-            # latest user in the full session and take a capped forward window.
-            latest_user = next(
-                (i for i in range(len(self.messages) - 1, -1, -1)
-                 if self.messages[i].get("role") == "user"),
-                None,
-            )
-            if latest_user is not None:
-                retained = self.messages[latest_user: latest_user + max_messages]
-
-        # Mirror get_history(): avoid persisting orphan tool results at the front.
-        start = find_legal_message_start(retained)
-        if start:
-            retained = retained[start:]
-
-        # Hard-cap guarantee unless the caller requested user-turn extension.
-        if not extend_to_user and len(retained) > max_messages:
-            retained = retained[-max_messages:]
-            start = find_legal_message_start(retained)
-            if start:
-                retained = retained[start:]
-
-        # Compute actually-dropped messages using identity comparison so that
-        # even when retained is a non-contiguous slice of original (the else
-        # branch above), we never duplicate or lose messages.
-        retained_ids = set(id(m) for m in retained)
-        dropped = [m for m in original if id(m) not in retained_ids]
-
-        # Count how many dropped messages were in the already-consolidated
-        # prefix of the original list.  This cannot be a simple min() because
-        # dropped may include messages from *after* the consolidated prefix
-        # (e.g. in the else branch).
-        already_consolidated = sum(
-            1 for i, m in enumerate(original)
-            if i < before_lc and id(m) not in retained_ids
-        )
-
-        # New last_archived = count of retained messages that were inside
-        # the old consolidated prefix.
-        new_lc = sum(
-            1 for i, m in enumerate(original)
-            if i < before_lc and id(m) in retained_ids
-        )
-
-        self.messages = retained
-        self.last_archived = new_lc
-        if dropped:
-            self.provider_state = None
-        self.updated_at = datetime.now()
-        return RetentionResult(
-            dropped=dropped,
-            already_consolidated_count=already_consolidated,
-        )
-
 class SessionPayload(TypedDict):
     key: str
     created_at: str | None
@@ -2010,7 +1903,7 @@ class SessionManager:
         user_index = 0
         found_target = False
         for message in source.messages:
-            if message.get("role") == "user":
+            if message.get("role") == "user" and not is_hidden_history_message(message):
                 if user_index == before_user_index:
                     found_target = True
                     break

nanobot/session/summary.py

@@ -3,15 +3,27 @@
 from __future__ import annotations
 
 from collections.abc import Mapping
+from dataclasses import dataclass
 from datetime import datetime
 from typing import TypedDict, cast
 
+SUMMARY_CONTINUATION_TEXT = (
+    "Continue the active task from the working-memory checkpoint above."
+)
 
 class SessionSummary(TypedDict):
     text: str
     last_active: str
 
 
+@dataclass(frozen=True, slots=True)
+class SessionSummaryCheckpoint:
+    """A replacement summary and the raw transcript boundary it covers."""
+
+    summary: str
+    transcript_boundary: int
+
+
 def session_summary_from_metadata(
     metadata: Mapping[str, object] | None,
     *,

nanobot/webui/settings_system.py

@@ -114,7 +114,6 @@ def system_settings_payload(
             "heartbeat": {
                 "enabled": config.gateway.heartbeat.enabled,
                 "interval_s": config.gateway.heartbeat.interval_s,
-                "keep_recent_messages": config.gateway.heartbeat.keep_recent_messages,
             },
             "dream": {
                 "schedule": defaults.dream.describe_schedule(),

tests/agent/test_auto_compact.py

@@ -8,7 +8,6 @@ from unittest.mock import AsyncMock, MagicMock
 import pytest
 
 from nanobot.agent.loop import AgentLoop
-from nanobot.agent.runner import AgentRunResult
 from nanobot.agent.tools.registry import ToolRegistry
 from nanobot.bus.events import InboundMessage
 from nanobot.bus.queue import MessageBus
@@ -229,36 +228,6 @@ class TestAgentLoopTTLParam:
         loop = _make_loop(tmp_path, session_ttl_minutes=0)
         assert loop.auto_compact._ttl == 0
 
-    @pytest.mark.asyncio
-    async def test_process_message_reads_history_with_token_budget(self, tmp_path):
-        """_process_message should pass an auto-derived token budget to get_history."""
-        loop = _make_loop(tmp_path)
-        session = loop.sessions.get_or_create("cli:direct")
-        session.get_history = MagicMock(return_value=[])
-        loop.context.build_messages = MagicMock(return_value=[])
-        loop._run_agent_loop = AsyncMock(
-            return_value=AgentRunResult(
-                final_content="ok",
-                messages=[],
-                stop_reason="stop",
-            )
-        )
-        loop._save_turn = MagicMock()
-
-        msg = InboundMessage(
-            channel="cli",
-            sender_id="u1",
-            chat_id="direct",
-            content="hello",
-        )
-        await loop._process_message(msg)
-        session.get_history.assert_called_once()
-        kwargs = session.get_history.call_args.kwargs
-        assert isinstance(kwargs.get("max_tokens"), int)
-        assert kwargs["max_tokens"] > 0
-        assert set(kwargs) == {"max_tokens", "extend_to_user"}
-
-
 class TestAutoCompact:
     """Test the _archive method."""
 

tests/agent/test_consolidator.py

@@ -55,9 +55,7 @@ def runtime(mock_provider):
 def consolidator(store):
     sessions = MagicMock()
     sessions.save = MagicMock()
-    # When maybe_consolidate_by_tokens refreshes the session reference via
-    # get_or_create(session.key), it should get back the same object the test
-    # passed in.  Store sessions by key so the lookup is transparent.
+    # Store sessions by key so refreshes observe the same test object.
     _session_cache: dict[str, MagicMock] = {}
     sessions.get_or_create = MagicMock(side_effect=lambda key: _session_cache.get(key, MagicMock()))
     sessions._session_cache = _session_cache
@@ -93,11 +91,17 @@ def _provider_state() -> ProviderConversationState:
 
 
 def _build_test_messages(**kwargs):
-    return [
-        {"role": "system", "content": "system prompt"},
+    system = "system prompt"
+    session_summary = kwargs.get("session_summary")
+    if session_summary:
+        system += f"\n\n[Archived Context Summary]\n{session_summary['text']}"
+    messages = [
+        {"role": "system", "content": system},
         *kwargs["history"],
-        {"role": "user", "content": kwargs["current_message"]},
     ]
+    if kwargs["current_message"] is not None:
+        messages.append({"role": "user", "content": kwargs["current_message"]})
+    return messages
 
 
 async def _archive(
@@ -112,15 +116,85 @@ async def _archive(
         messages,
         runtime=runtime,
         session_key=session_key,
-        request_messages=_build_test_messages(
-            history=messages,
-            current_message="consolidate",
-        ),
+        history=[
+            {"role": "system", "content": "system prompt"},
+            *messages,
+        ],
         request_tools=[],
         previous_summary=previous_summary,
     )
 
 
+class TestTurnTranscriptSummary:
+    async def test_uses_exact_accepted_prefix_and_existing_archiver(
+        self,
+        consolidator,
+        mock_provider,
+        runtime,
+    ):
+        accepted = [
+            {"role": "system", "content": "stable system"},
+            {"role": "user", "content": "accepted history"},
+        ]
+        tools = [{"type": "function", "function": {"name": "inspect"}}]
+        mock_provider.chat_with_retry.return_value = LLMResponse(
+            content="replacement checkpoint",
+        )
+
+        summary = await consolidator.summarize_transcript(
+            accepted,
+            "previous checkpoint",
+            runtime=runtime,
+            session_key="test:turn",
+            tools=tools,
+        )
+
+        assert summary == "replacement checkpoint"
+        call = mock_provider.chat_with_retry.await_args.kwargs
+        assert call["messages"][:-1] == accepted
+        assert call["messages"][-1]["role"] == "user"
+        assert "SNIP" in call["messages"][-1]["content"]
+        assert call["tools"] == tools
+
+    async def test_native_compaction_appends_only_archive_prompt(
+        self,
+        consolidator,
+        mock_provider,
+        runtime,
+    ):
+        accepted = [
+            {"role": "system", "content": "stable system"},
+            {"role": "user", "content": "raw history must not be replayed"},
+        ]
+        state = _provider_state()
+        mock_provider.can_resume_conversation_state.return_value = True
+        mock_provider.chat_with_retry.return_value = LLMResponse(
+            content="replacement checkpoint",
+        )
+
+        summary = await consolidator.summarize_provider_compaction(
+            state,
+            accepted,
+            "previous checkpoint",
+            runtime=runtime,
+            session_key="test:turn",
+            tools=[{"type": "function", "function": {"name": "inspect"}}],
+        )
+
+        assert summary == "replacement checkpoint"
+        call = mock_provider.chat_with_retry.await_args.kwargs
+        assert call["messages"][0] == accepted[0]
+        assert call["messages"][-1]["content"] == _ARCHIVE_PROMPT
+        assert accepted[1] not in call["messages"]
+        assert call["tools"] == []
+        provider_context = call["provider_context"]
+        assert provider_context.conversation_state is not None
+        assert provider_context.conversation_state.payload == state.payload
+        assert provider_context.conversation_state.pending_messages == [
+            call["messages"][-1],
+        ]
+
+
 class TestConsolidatorSummarize:
     def test_format_messages_keeps_media_only_user_turn(self):
         path = "/home/user/.nanobot/media/websocket/clip.mp4"
@@ -379,32 +453,7 @@ class TestConsolidatorArchiveErrorHandling:
         consolidator.store.raw_archive.assert_not_called()
 
 
-class TestConsolidatorTokenBudget:
-    async def test_prompt_below_threshold_does_not_consolidate(
-        self, consolidator, runtime
-    ):
-        """No consolidation when tokens are within budget."""
-        session = MagicMock()
-        session.last_archived = 0
-        session.messages = [{"role": "user", "content": "hi"}]
-        session.key = "test:key"
-        consolidator.sessions._session_cache[session.key] = session
-        consolidator.estimate_session_prompt_tokens = MagicMock(return_value=(100, "tiktoken"))
-        consolidator.archive_session = AsyncMock(return_value=True)
-        await consolidator.maybe_consolidate_by_tokens(session, runtime=runtime)
-        consolidator.archive_session.assert_not_called()
-
-    async def test_token_estimation_failure_propagates(self, consolidator, runtime):
-        session = Session(key="test:estimate-failure")
-        session.add_message("user", "hello")
-        consolidator.sessions._session_cache[session.key] = session
-        consolidator.estimate_session_prompt_tokens = MagicMock(
-            side_effect=RuntimeError("counter failed")
-        )
-
-        with pytest.raises(RuntimeError, match="counter failed"):
-            await consolidator.maybe_consolidate_by_tokens(session, runtime=runtime)
-
+class TestConsolidatorPromptEstimate:
     async def test_estimate_uses_full_unarchived_tail(self, consolidator, runtime):
         """Consolidation pressure must account for the full unarchived tail."""
         session = Session(key="test:full-tail")
@@ -443,129 +492,6 @@ class TestConsolidatorTokenBudget:
         assert len(captured["history"]) == 8
         assert captured["history"][0]["content"] == "msg-2"
 
-    async def test_token_overflow_appends_prompt_to_replay_prefix(
-        self,
-        consolidator,
-        mock_provider,
-        runtime,
-    ):
-        consolidator._SAFETY_BUFFER = 0
-        session = Session(key="test:token-prefix")
-        session.provider_state = _provider_state()
-        session.messages = [
-            {
-                "role": "user" if i in {0, 50, 61} else "assistant",
-                "content": f"m{i}",
-            }
-            for i in range(70)
-        ]
-        consolidator.sessions._session_cache[session.key] = session
-        consolidator.estimate_session_prompt_tokens = MagicMock(
-            side_effect=[(1200, "tiktoken"), (400, "tiktoken")]
-        )
-        consolidator.pick_consolidation_boundary = MagicMock(return_value=50)
-        consolidator.archiver._build_messages = MagicMock(side_effect=_build_test_messages)
-        mock_provider.estimate_prompt_tokens.return_value = (100, "test-counter")
-        mock_provider.chat_with_retry.return_value = LLMResponse(
-            content="Token overflow summary.",
-            finish_reason="stop",
-        )
-
-        await consolidator.maybe_consolidate_by_tokens(session, runtime=runtime)
-
-        request = mock_provider.chat_with_retry.await_args.kwargs
-        assert [message["content"] for message in request["messages"][1:-1]] == [
-            f"m{i}" for i in range(50)
-        ]
-        assert request["messages"][-1]["content"] == _ARCHIVE_PROMPT
-        assert request["tools"] == []
-        assert "tool_choice" not in request
-        assert session.last_archived == 50
-        assert session.provider_state == _provider_state()
-
-    async def test_raw_archive_fallback_advances_archive_watermark(
-        self, consolidator, runtime
-    ):
-        """When archive() falls back to raw-archive (LLM failed), the cursor
-        must still advance. Otherwise the same chunk gets raw-archived again
-        on every subsequent maybe_consolidate_by_tokens() call, spamming
-        duplicate [RAW] entries into history.jsonl."""
-        consolidator._SAFETY_BUFFER = 0
-        session = Session(key="test:key")
-        session.provider_state = _provider_state()
-        session.messages = [
-            {"role": "user" if i in {0, 50} else "assistant", "content": f"m{i}"}
-            for i in range(70)
-        ]
-        consolidator.sessions._session_cache[session.key] = session
-        consolidator.estimate_session_prompt_tokens = MagicMock(
-            side_effect=[(1200, "tiktoken"), (400, "tiktoken")]
-        )
-        consolidator.archive_session = AsyncMock(return_value="[RAW] checkpoint")
-
-        await consolidator.maybe_consolidate_by_tokens(session, runtime=runtime)
-
-        consolidator.archive_session.assert_awaited_once()
-        # The chunk is considered "materialized" (as a raw-archive breadcrumb),
-        # so the archive watermark must have moved past it without touching
-        # the provider-owned continuation state.
-        assert session.last_archived == 50
-        assert session.provider_state == _provider_state()
-
-    async def test_raw_archive_fallback_breaks_round_loop(
-        self, consolidator, runtime
-    ):
-        """A degraded LLM should not trigger more archive() calls within the
-        same maybe_consolidate_by_tokens invocation — bail after one fallback."""
-        consolidator._SAFETY_BUFFER = 0
-        session = MagicMock()
-        session.last_archived = 0
-        session.key = "test:key"
-        session.messages = [
-            {"role": "user" if i in {0, 20, 40, 60} else "assistant", "content": f"m{i}"}
-            for i in range(70)
-        ]
-        session.metadata = {}
-        consolidator.sessions._session_cache[session.key] = session
-        # Keep estimates high so the loop would otherwise run multiple rounds.
-        consolidator.estimate_session_prompt_tokens = MagicMock(
-            return_value=(1200, "tiktoken")
-        )
-        consolidator.archive_session = AsyncMock(return_value="[RAW] checkpoint")
-
-        await consolidator.maybe_consolidate_by_tokens(session, runtime=runtime)
-
-        # The fixed policy archives at most one prefix per call.
-        assert consolidator.archive_session.await_count == 1
-
-    async def test_boundary_respected_when_no_intermediate_user_turn(
-        self, consolidator, runtime
-    ):
-        """When boundary points past a long tool chain, the full chunk is archived."""
-        consolidator._SAFETY_BUFFER = 0
-        session = MagicMock()
-        session.last_archived = 0
-        session.key = "test:key"
-        session.messages = [
-            {
-                "role": "user" if i in {0, 61} else "assistant",
-                "content": f"m{i}",
-            }
-            for i in range(70)
-        ]
-        consolidator.sessions._session_cache[session.key] = session
-        consolidator.estimate_session_prompt_tokens = MagicMock(
-            side_effect=[(1200, "tiktoken"), (400, "tiktoken")]
-        )
-        consolidator.archive_session = AsyncMock(return_value=True)
-
-        await consolidator.maybe_consolidate_by_tokens(session, runtime=runtime)
-
-        consolidator.archive_session.assert_awaited_once()
-        # The fixed recent tail expands backward to the user at idx=61.
-        assert session.last_archived == 61
-
-
 class TestCompactIdleSession:
     """Idle compaction tests."""
 
@@ -1347,110 +1273,6 @@ class TestCompactIdleSession:
         assert not lock.locked()
 
 
-class TestConsolidatorSessionRefresh:
-    """Background consolidation must detect stale session references."""
-
-    @pytest.mark.asyncio
-    async def test_reloads_before_empty_session_guard(self, tmp_path):
-        """A stale empty reference must not skip a non-empty cached session."""
-        from nanobot.agent.memory import Consolidator, MemoryStore
-        from nanobot.session.manager import Session, SessionManager
-
-        store = MemoryStore(tmp_path)
-        provider = MagicMock()
-        provider.chat_with_retry = AsyncMock(
-            return_value=MagicMock(content="summary", finish_reason="stop")
-        )
-        provider.generation = GenerationSettings(max_tokens=4096)
-        provider.estimate_prompt_tokens = MagicMock(return_value=(10, "test"))
-        runtime = LLMRuntime.capture(
-            provider,
-            "test-model",
-            context_window_tokens=128_000,
-        )
-        sessions = SessionManager(tmp_path)
-        consolidator = Consolidator(
-            store=store,
-            sessions=sessions,
-            build_messages=MagicMock(return_value=[]),
-            get_tool_definitions=MagicMock(return_value=[]),
-        )
-
-        fresh = sessions.get_or_create("cli:test")
-        fresh.add_message("user", "fresh message")
-        sessions.save(fresh)
-        stale_empty = Session(key="cli:test")
-
-        seen: dict[str, Session] = {}
-
-        def estimate(session: Session, *, runtime):
-            seen["session"] = session
-            return 10, "test"
-
-        consolidator.estimate_session_prompt_tokens = MagicMock(side_effect=estimate)
-
-        await consolidator.maybe_consolidate_by_tokens(
-            stale_empty,
-            runtime=runtime,
-        )
-
-        assert seen["session"] is fresh
-
-    @pytest.mark.asyncio
-    async def test_reloads_stale_session_after_compact(self, tmp_path):
-        """After compact_idle_session replaces the session, a concurrent
-        maybe_consolidate_by_tokens with the old reference should use the
-        fresh session from cache instead of overwriting."""
-        from nanobot.agent.memory import Consolidator, MemoryStore
-        from nanobot.session.manager import SessionManager
-
-        store = MemoryStore(tmp_path)
-        provider = MagicMock()
-        provider.chat_with_retry = AsyncMock(
-            return_value=MagicMock(content="summary", finish_reason="stop")
-        )
-        provider.generation = GenerationSettings(max_tokens=4096)
-        provider.estimate_prompt_tokens = MagicMock(return_value=(10, "test"))
-        runtime = LLMRuntime.capture(
-            provider,
-            "test-model",
-            context_window_tokens=128_000,
-        )
-        sessions = SessionManager(tmp_path)
-        consolidator = Consolidator(
-            store=store,
-            sessions=sessions,
-            build_messages=MagicMock(return_value=[]),
-            get_tool_definitions=MagicMock(return_value=[]),
-        )
-
-        # Populate session with many messages
-        session = sessions.get_or_create("cli:test")
-        for i in range(20):
-            session.add_message("user", f"u{i}")
-            session.add_message("assistant", f"a{i}")
-        sessions.save(session)
-
-        # Simulate: background consolidation captures old reference
-        old_ref = session
-
-        await consolidator.compact_idle_session(
-            "cli:test",
-            runtime=runtime,
-            max_suffix=8,
-        )
-
-        await consolidator.maybe_consolidate_by_tokens(
-            old_ref,
-            runtime=runtime,
-        )
-
-        session_after = sessions.get_or_create("cli:test")
-        assert len(session_after.messages) == 40
-        assert session_after.last_archived == 40
-        assert len(session_after.get_history(max_messages=40)) == 8
-
-
 class TestRawArchiveTruncation:
     """raw_archive() must cap entry size to avoid bloating history.jsonl."""
 

tests/agent/test_dream.py

@@ -404,10 +404,9 @@ class TestEphemeralDirect:
         with (
             patch("nanobot.agent.loop.SessionManager"),
             patch("nanobot.agent.loop.SubagentManager") as mock_sub,
-            patch("nanobot.agent.loop.Consolidator") as mock_consolidator_cls,
+            patch("nanobot.agent.loop.Consolidator"),
         ):
             mock_sub.return_value.cancel_by_session = AsyncMock(return_value=0)
-            mock_consolidator_cls.return_value.maybe_consolidate_by_tokens = AsyncMock()
             loop = AgentLoop(
                 bus=bus,
                 provider=provider,
@@ -493,20 +492,6 @@ class TestEphemeralDirect:
 
         assert captured.get("ephemeral") is False
 
-    async def test_ephemeral_skips_consolidator(self, tmp_path, _make_loop):
-        """When ephemeral=True, consolidator.maybe_consolidate_by_tokens is not called."""
-        from unittest.mock import patch
-
-        loop, store = _make_loop
-
-        with patch.object(
-            loop.consolidator, "maybe_consolidate_by_tokens",
-        ) as mock_consolidate:
-            await loop.process_direct(
-                "test", session_key="dream:consolidate-test", ephemeral=True,
-            )
-            mock_consolidate.assert_not_called()
-
     async def test_ephemeral_response_reports_stop_reason(self, tmp_path, _make_loop):
         loop, store = _make_loop
         loop.provider.chat_with_retry.return_value = LLMResponse(
@@ -701,10 +686,9 @@ class TestEphemeralHooks:
         with (
             patch("nanobot.agent.loop.SessionManager"),
             patch("nanobot.agent.loop.SubagentManager") as mock_sub,
-            patch("nanobot.agent.loop.Consolidator") as mock_consolidator_cls,
+            patch("nanobot.agent.loop.Consolidator"),
         ):
             mock_sub.return_value.cancel_by_session = AsyncMock(return_value=0)
-            mock_consolidator_cls.return_value.maybe_consolidate_by_tokens = AsyncMock()
             loop = AgentLoop(
                 bus=bus,
                 provider=provider,

tests/agent/test_history_replay.py

@@ -12,6 +12,7 @@ from nanobot.bus.events import InboundMessage
 from nanobot.bus.queue import MessageBus
 from nanobot.providers.base import LLMResponse
 from nanobot.session.manager import Session
+from nanobot.session.summary import SUMMARY_CONTINUATION_TEXT
 
 
 def _make_loop(tmp_path: Path, context_window_tokens: int = 200_000) -> AgentLoop:
@@ -66,13 +67,12 @@ def test_explicit_message_limit_still_starts_at_user_turn() -> None:
 
 
 @pytest.mark.asyncio
-async def test_process_message_replays_with_token_budget_only(tmp_path: Path) -> None:
+async def test_process_message_hands_complete_replay_to_runner(tmp_path: Path) -> None:
     loop = _make_loop(tmp_path, context_window_tokens=32_768)
     loop.provider.chat_with_retry = AsyncMock(
         return_value=LLMResponse(content="ok", tool_calls=[], usage=None)
     )
     loop.tools.get_definitions = MagicMock(return_value=[])
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     session = loop.sessions.get_or_create("cli:test")
     with patch.object(session, "get_history", wraps=session.get_history) as get_history:
@@ -81,20 +81,16 @@ async def test_process_message_replays_with_token_budget_only(tmp_path: Path) ->
         )
 
     assert result is not None
-    assert get_history.call_args.kwargs == {
-        "max_tokens": loop._replay_token_budget(loop.llm_runtime()),
-        "extend_to_user": False,
-    }
+    assert get_history.call_args.kwargs == {"extend_to_user": False}
 
 
 @pytest.mark.asyncio
-async def test_token_budget_keeps_current_user_as_replay_boundary(tmp_path: Path) -> None:
+async def test_runner_checkpoint_keeps_current_user_as_replay_boundary(tmp_path: Path) -> None:
     loop = _make_loop(tmp_path, context_window_tokens=8_000)
     loop.provider.chat_with_retry = AsyncMock(
         return_value=LLMResponse(content="ok", tool_calls=[], usage=None)
     )
     loop.tools.get_definitions = MagicMock(return_value=[])
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     session = loop.sessions.get_or_create("cli:test")
     session.add_message("user", "old")
@@ -117,4 +113,6 @@ async def test_token_budget_keeps_current_user_as_replay_boundary(tmp_path: Path
     sent_messages = loop.provider.chat_with_retry.await_args.kwargs["messages"]
     sent_text = "\n".join(str(message.get("content")) for message in sent_messages)
     assert "new question" in sent_text
-    assert "long older turn" not in sent_text
+    assert [message["role"] for message in sent_messages] == ["system", "user", "user"]
+    assert sent_messages[1]["content"] == SUMMARY_CONTINUATION_TEXT
+    assert any(message.get("content") == "long older turn" for message in session.messages)

tests/agent/test_loop_consolidation_tokens.py

@@ -4,7 +4,12 @@ import pytest
 
 from nanobot.agent.loop import AgentLoop
 from nanobot.bus.queue import MessageBus
-from nanobot.providers.base import LLMResponse
+from nanobot.providers.base import (
+    GenerationSettings,
+    LLMResponse,
+    ProviderConversationState,
+)
+from nanobot.session.summary import SUMMARY_CONTINUATION_TEXT
 
 
 def _make_loop(
@@ -14,7 +19,6 @@ def _make_loop(
     context_window_tokens: int,
     max_tokens: int = 0,
 ) -> AgentLoop:
-    from nanobot.providers.base import GenerationSettings
     provider = MagicMock()
     provider.get_default_model.return_value = "test-model"
     provider.generation = GenerationSettings(max_tokens=max_tokens)
@@ -39,186 +43,108 @@ def _make_loop(
 
 
 @pytest.mark.asyncio
-async def test_prompt_below_threshold_does_not_consolidate(tmp_path) -> None:
-    loop = _make_loop(tmp_path, estimated_tokens=100, context_window_tokens=200)
-    loop.consolidator.archive_session = AsyncMock(return_value=True)  # type: ignore[method-assign]
-
-    await loop.process_direct("hello", session_key="cli:test")
-
-    loop.consolidator.archive_session.assert_not_awaited()
-
-
-@pytest.mark.asyncio
-async def test_prompt_above_threshold_triggers_consolidation(tmp_path) -> None:
-    loop = _make_loop(tmp_path, estimated_tokens=1000, context_window_tokens=200)
-    loop.consolidator.archive_session = AsyncMock(return_value=True)  # type: ignore[method-assign]
-    session = loop.sessions.get_or_create("cli:test")
-    session.messages = [
-        {"role": role, "content": f"{role[0]}{turn}"}
-        for turn in range(10)
-        for role in ("user", "assistant")
-    ]
-    loop.sessions.save(session)
-
-    await loop.process_direct("hello", session_key="cli:test")
-
-    assert loop.consolidator.archive_session.await_count >= 1
-
-
-@pytest.mark.asyncio
-async def test_token_consolidation_refreshes_summary_for_current_request(tmp_path) -> None:
-    loop = _make_loop(tmp_path, estimated_tokens=0, context_window_tokens=200)
-    loop.consolidator.archive_session = AsyncMock(  # type: ignore[method-assign]
-        return_value="FRESH_CHECKPOINT"
-    )
-    loop.consolidator.estimate_session_prompt_tokens = MagicMock(  # type: ignore[method-assign]
-        return_value=(1000, "test")
-    )
+async def test_runner_pressure_commits_summary_and_current_delta(tmp_path) -> None:
+    loop = _make_loop(tmp_path, estimated_tokens=100, context_window_tokens=2_000)
+    loop.context_block_limit = 500
+    loop.provider.generation = GenerationSettings(max_tokens=100)
+    loop.provider.can_resume_conversation_state.return_value = False
     loop.schedule_background = lambda coro: coro.close()  # type: ignore[method-assign]
 
     session = loop.sessions.get_or_create("cli:test")
     session.messages = [
-        {"role": role, "content": f"{role[0]}{turn}"}
-        for turn in range(10)
+        {"role": role, "content": f"old-{role}-{turn}"}
+        for turn in range(6)
         for role in ("user", "assistant")
     ]
     loop.sessions.save(session)
 
-    await loop.process_direct("hello", session_key="cli:test")
+    def estimate(messages, _tools, _model):
+        contents = [str(message.get("content")) for message in messages]
+        if contents and "SNIP" in contents[-1]:
+            return 300, "test-counter"
+        if any(content.startswith("old-") for content in contents):
+            return 600, "test-counter"
+        return 100, "test-counter"
 
-    request_messages = loop.provider.chat_with_retry.await_args.kwargs["messages"]
-    system_prompt = request_messages[0]["content"]
-    assert "FRESH_CHECKPOINT" in system_prompt
-    assert all(message.get("content") != "u0" for message in request_messages)
-    assert loop.sessions.get_or_create("cli:test").last_archived == 12
+    loop.provider.estimate_prompt_tokens.side_effect = estimate
+    loop.provider.chat_with_retry = AsyncMock(side_effect=[
+        LLMResponse(content="Current checkpoint.", tool_calls=[]),
+        LLMResponse(content="done", tool_calls=[]),
+    ])
 
+    result = await loop.process_direct("continue the task", session_key="cli:test")
 
-@pytest.mark.asyncio
-async def test_prompt_above_threshold_uses_fixed_recent_tail(tmp_path) -> None:
-    loop = _make_loop(tmp_path, estimated_tokens=1000, context_window_tokens=200)
-    loop.consolidator.archive_session = AsyncMock(return_value=True)  # type: ignore[method-assign]
-
-    session = loop.sessions.get_or_create("cli:test")
-    session.messages = [
-        {"role": role, "content": f"{role[0]}{turn}"}
-        for turn in range(10)
-        for role in ("user", "assistant")
-    ]
-    loop.sessions.save(session)
-
-    await loop.consolidator.maybe_consolidate_by_tokens(
-        session,
-        runtime=loop.llm_runtime(),
-    )
-
-    archive_end = loop.consolidator.archive_session.await_args.kwargs["archive_end"]
-    archived_chunk = session.messages[:archive_end]
-    assert [message["content"] for message in archived_chunk] == [
-        "u0", "a0", "u1", "a1", "u2", "a2", "u3", "a3", "u4", "a4", "u5", "a5",
-    ]
-    assert session.last_archived == 12
-
-
-@pytest.mark.asyncio
-async def test_consolidation_persists_summary_for_next_prepare_session(tmp_path) -> None:
-    loop = _make_loop(tmp_path, estimated_tokens=0, context_window_tokens=200)
-    loop.consolidator.archive_session = AsyncMock(return_value="User discussed project status.")  # type: ignore[method-assign]
-
-    session = loop.sessions.get_or_create("cli:test")
-    session.messages = [
-        {"role": role, "content": f"{role[0]}{turn}"}
-        for turn in range(5)
-        for role in ("user", "assistant")
-    ]
-    loop.sessions.save(session)
-
-    def mock_estimate(_session, *, runtime):
-        return (500, "test")
-
-    loop.consolidator.estimate_session_prompt_tokens = mock_estimate  # type: ignore[method-assign]
-
-    await loop.consolidator.maybe_consolidate_by_tokens(
-        session,
-        runtime=loop.llm_runtime(),
-    )
+    assert result.content == "done"
+    assert loop.provider.chat_with_retry.await_count == 2
+    model_request = loop.provider.chat_with_retry.await_args_list[1].kwargs["messages"]
+    assert "Current checkpoint." in model_request[0]["content"]
+    assert model_request[1]["content"] == SUMMARY_CONTINUATION_TEXT
+    assert model_request[2]["content"] == "continue the task"
 
     reloaded = loop.sessions.get_or_create("cli:test")
-    meta = reloaded.metadata.get("_last_summary")
-    assert meta is not None
-    assert meta["text"] == "User discussed project status."
-
-    reloaded, pending = loop.auto_compact.prepare_session(reloaded, "cli:test")
-    assert pending is not None
-    assert pending["text"] == "User discussed project status."
-    # _last_summary persists for restart survival.
-    assert "_last_summary" in reloaded.metadata
+    assert reloaded.messages[0]["content"] == "old-user-0"
+    assert reloaded.metadata["_last_summary"]["text"] == "Current checkpoint."
+    assert reloaded.messages[reloaded.last_archived]["content"] == (
+        SUMMARY_CONTINUATION_TEXT
+    )
+    assert [message["content"] for message in reloaded.get_history()] == [
+        SUMMARY_CONTINUATION_TEXT,
+        "continue the task",
+        "done",
+    ]
 
 
 @pytest.mark.asyncio
-async def test_preflight_consolidation_receives_pending_summary(tmp_path) -> None:
-    loop = _make_loop(tmp_path, estimated_tokens=100, context_window_tokens=200)
-    session = loop.sessions.get_or_create("cli:test")
-    loop.auto_compact.prepare_session = MagicMock(
-        return_value=(
-            session,
-            {"text": "earlier context", "last_active": session.updated_at.isoformat()},
-        )
-    )  # type: ignore[method-assign]
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=None)  # type: ignore[method-assign]
-    loop.schedule_background = lambda coro: coro.close()  # type: ignore[method-assign]
-
-    runtime = loop.llm_runtime()
-    await loop.process_direct("hello", session_key="cli:test", runtime=runtime)
-
-    loop.consolidator.maybe_consolidate_by_tokens.assert_any_await(
-        session,
-        runtime=runtime,
-    )
-    assert len(loop.consolidator.maybe_consolidate_by_tokens.call_args_list) == 2
-    assert all(
-        call.kwargs["runtime"] is runtime
-        for call in loop.consolidator.maybe_consolidate_by_tokens.call_args_list
-    )
-
-
-@pytest.mark.asyncio
-async def test_preflight_consolidation_before_llm_call(tmp_path) -> None:
-    """Verify preflight consolidation runs before the LLM call in process_direct."""
-    order: list[str] = []
-
-    loop = _make_loop(tmp_path, estimated_tokens=0, context_window_tokens=200)
-
-    archived_session_keys: list[str | None] = []
-
-    async def track_consolidate(session, *, archive_end, runtime):
-        order.append("consolidate")
-        archived_session_keys.append(session.key)
-        return True
-    loop.consolidator.archive_session = track_consolidate  # type: ignore[method-assign]
-
-    async def track_llm(*args, **kwargs):
-        order.append("llm")
-        return LLMResponse(content="ok", tool_calls=[])
-    loop.provider.chat_with_retry = track_llm
-    loop.provider.chat_stream_with_retry = track_llm
-    loop.schedule_background = lambda coro: coro.close()  # type: ignore[method-assign]
-
-    session = loop.sessions.get_or_create("cli:test")
+async def test_native_provider_compaction_commits_portable_terminal_checkpoint(
+    tmp_path,
+) -> None:
+    loop = _make_loop(tmp_path, estimated_tokens=100, context_window_tokens=2_000)
+    session = loop.sessions.get_or_create("cli:native")
     session.messages = [
-        {"role": role, "content": f"{role[0]}{turn}"}
-        for turn in range(10)
-        for role in ("user", "assistant")
+        {"role": "user", "content": "accepted history"},
+        {"role": "assistant", "content": "accepted answer"},
     ]
     loop.sessions.save(session)
-    call_count = [0]
-    def mock_estimate(_session, *, runtime):
-        call_count[0] += 1
-        return (1000 if call_count[0] <= 1 else 80, "test")
-    loop.consolidator.estimate_session_prompt_tokens = mock_estimate  # type: ignore[method-assign]
+    compacted_state = ProviderConversationState(
+        kind="openai_responses",
+        provider="openai:test",
+        model="test-model",
+        version=1,
+        payload={"items": [{"type": "compaction", "encrypted_content": "opaque"}]},
+    )
+    loop.provider.can_resume_conversation_state.return_value = True
+    loop.provider.chat_with_retry = AsyncMock(return_value=LLMResponse(
+        content="done",
+        provider_state=compacted_state,
+        provider_compaction_applied=True,
+        provider_compaction_state=compacted_state,
+        provider_compaction_scope="current_request",
+    ))
+    loop.consolidator.summarize_provider_compaction = AsyncMock(
+        return_value="portable terminal checkpoint",
+    )
 
-    await loop.process_direct("hello", session_key="cli:test")
+    result = await loop.process_direct("continue", session_key="cli:native")
 
-    assert "consolidate" in order
-    assert "llm" in order
-    assert order.index("consolidate") < order.index("llm")
-    assert archived_session_keys == ["cli:test"]
+    assert result.content == "done"
+    summarize = loop.consolidator.summarize_provider_compaction
+    summarize.assert_awaited_once()
+    assert summarize.await_args.args[0] == compacted_state
+    accepted = summarize.await_args.args[1]
+    accepted_contents = [message.get("content") for message in accepted]
+    assert "accepted history" in accepted_contents
+    assert "accepted answer" in accepted_contents
+    assert "continue" in accepted_contents
+    assert "done" not in accepted_contents
+    reloaded = loop.sessions.get_or_create("cli:native")
+    assert reloaded.provider_state is None
+    assert reloaded.metadata["_last_summary"]["text"] == (
+        "portable terminal checkpoint"
+    )
+    assert reloaded.messages[reloaded.last_archived]["content"] == (
+        SUMMARY_CONTINUATION_TEXT
+    )
+    assert [message["content"] for message in reloaded.get_history()] == [
+        SUMMARY_CONTINUATION_TEXT,
+        "done",
+    ]

tests/agent/test_loop_image_generation_media.py

@@ -69,7 +69,6 @@ async def test_outbound_no_longer_carries_generated_media(
         ),
         image_generation_provider_config=ProviderConfig(api_key="sk-or-test"),
     )
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     result = await loop._process_message(
         InboundMessage(

tests/agent/test_loop_progress.py

@@ -425,7 +425,6 @@ class TestToolEventProgress:
                 None,
             ),
         )
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="websocket",
@@ -473,7 +472,6 @@ class TestToolEventProgress:
         provider.chat_stream_with_retry = AsyncMock()
         loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="openai-codex/gpt-5.5")
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="whatsapp",
@@ -512,7 +510,6 @@ class TestToolEventProgress:
         loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="openai-codex/gpt-5.5")
         _attach_webui_runtime_events(loop, bus)
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="websocket",
@@ -566,7 +563,6 @@ class TestToolEventProgress:
         loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="test-model")
         _attach_webui_runtime_events(loop, bus)
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="websocket",
@@ -611,7 +607,6 @@ class TestToolEventProgress:
         loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="test-model")
         _attach_webui_runtime_events(loop, bus)
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="websocket",
@@ -655,7 +650,6 @@ class TestToolEventProgress:
         _attach_webui_runtime_events(loop, bus)
         loop.max_iterations = 1
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="websocket",
@@ -747,7 +741,6 @@ class TestToolEventProgress:
         )
         _attach_webui_runtime_events(loop, bus)
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="websocket",
@@ -815,9 +808,6 @@ class TestToolEventProgress:
             return "ok"
 
         loop.tools.execute = execute_tool
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(  # type: ignore[method-assign]
-            return_value=False
-        )
 
         session_key = "websocket:chat-a"
         session = loop.sessions.get_or_create(session_key)
@@ -949,7 +939,6 @@ class TestToolEventProgress:
         loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="openai-codex/gpt-5.5")
         _attach_webui_runtime_events(loop, bus)
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="websocket",
@@ -1048,7 +1037,6 @@ class TestToolEventProgress:
         loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="test-model")
         _attach_webui_runtime_events(loop, bus)
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="websocket",
@@ -1132,7 +1120,6 @@ class TestToolEventProgress:
         loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="test-model")
         _attach_webui_runtime_events(loop, bus)
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await asyncio.wait_for(loop._dispatch(InboundMessage(
             channel="websocket",
@@ -1181,7 +1168,6 @@ class TestToolEventProgress:
         loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="test-model")
         _attach_webui_runtime_events(loop, bus)
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         captured: dict[str, object] = {}
 
@@ -1268,7 +1254,6 @@ class TestToolEventProgress:
         provider.chat_with_retry = AsyncMock(return_value=LLMResponse(content="Done", tool_calls=[]))
         loop = AgentLoop(bus=bus, provider=provider, workspace=tmp_path, model="test-model")
         loop.tools.get_definitions = MagicMock(return_value=[])
-        loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
         await loop._dispatch(InboundMessage(
             channel="slack",

tests/agent/test_loop_runner_integration.py

@@ -112,7 +112,6 @@ async def test_goal_command_can_implement_plan_from_prior_discussion(tmp_path):
         LLMResponse(content="done", tool_calls=[], usage=None),
     ])
     loop = AgentLoop(bus=MessageBus(), provider=provider, workspace=tmp_path, model="test-model")
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=None)
     session = loop.sessions.get_or_create("cli:direct")
     session.add_message("user", "Let's agree on the migration implementation.")
     session.add_message("assistant", "Use the staged migration plan and run integration tests.")
@@ -166,7 +165,6 @@ async def test_runtime_context_is_persisted_as_next_turn_prompt_prefix(tmp_path)
         LLMResponse(content="second answer", usage=None),
     ])
     loop = AgentLoop(bus=MessageBus(), provider=provider, workspace=tmp_path, model="test-model")
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=None)
     session = loop.sessions.get_or_create("cli:direct")
     provider_calls: list[str | None] = []
 
@@ -220,7 +218,6 @@ async def test_webui_quote_reaches_model_without_leaking_into_public_history(tmp
     provider.generation = GenerationSettings()
     provider.chat_with_retry = AsyncMock(return_value=LLMResponse(content="answer", usage=None))
     loop = AgentLoop(bus=MessageBus(), provider=provider, workspace=tmp_path, model="test-model")
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=None)
     session = loop.sessions.get_or_create("websocket:chat")
     quote = webui_quote_runtime_context({
         WEBUI_QUOTE_METADATA: "the selected answer excerpt",
@@ -265,7 +262,6 @@ async def test_runtime_context_provider_runs_once_across_tool_iterations(tmp_pat
         LLMResponse(content="done", usage=None),
     ])
     loop = AgentLoop(bus=MessageBus(), provider=provider, workspace=tmp_path, model="test-model")
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=None)
     provider_calls = 0
 
     async def provide_context(_request):
@@ -310,7 +306,6 @@ async def test_non_goal_direct_turn_cannot_reuse_prior_goal_command(tmp_path):
         LLMResponse(content="handled as a one-time task", tool_calls=[], usage=None),
     ])
     loop = AgentLoop(bus=MessageBus(), provider=provider, workspace=tmp_path, model="test-model")
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=None)
     session = loop.sessions.get_or_create("api:default")
     session.add_message("user", "/goal old completed request")
     session.add_message("assistant", "The old request is complete.")
@@ -589,7 +584,6 @@ async def test_next_turn_after_llm_error_keeps_turn_boundary(tmp_path):
 
     loop = AgentLoop(bus=MessageBus(), provider=provider, workspace=tmp_path, model="test-model")
     loop.tools.get_definitions = MagicMock(return_value=[])
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     first = await loop._process_message(
         InboundMessage(channel="cli", sender_id="user", chat_id="test", content="first question")

tests/agent/test_loop_save_turn.py

@@ -45,6 +45,10 @@ from nanobot.session.recovery import (
     RUNTIME_CHECKPOINT_KEY,
     restore_runtime_checkpoint,
 )
+from nanobot.session.summary import (
+    SUMMARY_CONTINUATION_TEXT,
+    SessionSummaryCheckpoint,
+)
 from nanobot.session.turn_continuation import (
     INTERNAL_CONTINUATION_META,
     INTERNAL_CONTINUATION_RUN_STARTED_AT_META,
@@ -506,6 +510,60 @@ def test_save_turn_keeps_multimodal_runtime_context_for_model_replay() -> None:
     assert public_history_message(session.messages[0])["content"] == []
 
 
+def test_save_turn_commits_summary_boundary_without_rewriting_raw_history() -> None:
+    loop = _mk_loop()
+    session = Session(key="test:summary-checkpoint")
+    session.add_message("user", "inspect the project")
+    messages = [
+        {"role": "system", "content": "system"},
+        {"role": "user", "content": "inspect the project"},
+        {
+            "role": "assistant",
+            "content": "",
+            "tool_calls": [{
+                "id": "call-1",
+                "type": "function",
+                "function": {"name": "inspect", "arguments": "{}"},
+            }],
+        },
+        {
+            "role": "tool",
+            "tool_call_id": "call-1",
+            "name": "inspect",
+            "content": "full current result",
+        },
+        {"role": "assistant", "content": "done"},
+    ]
+
+    loop._save_turn(
+        session,
+        messages,
+        skip=2,
+        summary_checkpoint=SessionSummaryCheckpoint(
+            summary="Current working-memory checkpoint.",
+            transcript_boundary=2,
+        ),
+        input_persisted_early=True,
+    )
+
+    assert [message["role"] for message in session.messages] == [
+        "user", "user", "assistant", "tool", "assistant",
+    ]
+    assert session.messages[0]["content"] == "inspect the project"
+    assert session.messages[1]["content"] == SUMMARY_CONTINUATION_TEXT
+    assert session.messages[1]["_hidden_history"] is True
+    assert session.last_archived == 1
+    assert session.metadata["_last_summary"]["text"] == (
+        "Current working-memory checkpoint."
+    )
+    assert [message["content"] for message in session.get_history()] == [
+        SUMMARY_CONTINUATION_TEXT,
+        "",
+        "full current result",
+        "done",
+    ]
+
+
 def test_save_turn_acknowledges_every_merged_recovery_followup() -> None:
     """Persisting a merged injected row retires every durable follow-up ID."""
     loop = _mk_loop()
@@ -966,7 +1024,6 @@ async def test_runtime_checkpoint_keeps_provider_state_out_of_public_metadata(
 @pytest.mark.asyncio
 async def test_process_message_persists_user_message_before_turn_completes(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop._run_agent_loop = AsyncMock(side_effect=RuntimeError("boom"))  # type: ignore[method-assign]
 
     msg = InboundMessage(channel="feishu", sender_id="u1", chat_id="c1", content="persist me")
@@ -986,7 +1043,6 @@ async def test_subagent_followup_stages_provider_state_before_turn_runs(
     tmp_path: Path,
 ) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop._run_agent_loop = AsyncMock(side_effect=RuntimeError("boom"))  # type: ignore[method-assign]
     loop.provider.can_resume_conversation_state.return_value = True
     session = loop.sessions.get_or_create("cli:subagent-crash")
@@ -1016,7 +1072,6 @@ async def test_subagent_followup_state_is_durable_before_prompt_assembly(
     tmp_path: Path,
 ) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop.provider.can_resume_conversation_state.return_value = True
     loop.context.build_system_prompt = MagicMock(  # type: ignore[method-assign]
         side_effect=RuntimeError("prompt boom"),
@@ -1049,7 +1104,6 @@ async def test_subagent_redelivery_does_not_duplicate_staged_provider_input(
     tmp_path: Path,
 ) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop.provider.can_resume_conversation_state.return_value = True
     build_system_prompt = loop.context.build_system_prompt
     loop.context.build_system_prompt = MagicMock(  # type: ignore[method-assign]
@@ -1101,7 +1155,6 @@ async def test_subagent_followup_clears_state_before_compatibility_failure(
     tmp_path: Path,
 ) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop.provider.can_resume_conversation_state.side_effect = RuntimeError(
         "compatibility boom"
     )
@@ -1129,7 +1182,6 @@ async def test_subagent_followup_clears_state_before_compatibility_failure(
 async def test_process_message_persists_unified_session_delivery_route(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
     loop._unified_session = True
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop._run_agent_loop = AsyncMock(side_effect=RuntimeError("boom"))  # type: ignore[method-assign]
 
     msg = InboundMessage(
@@ -1230,7 +1282,6 @@ async def test_process_message_persists_media_paths_on_user_turn(tmp_path: Path)
     img_b.write_bytes(_PNG_1X1)
 
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop._run_agent_loop = AsyncMock(side_effect=RuntimeError("interrupt"))  # type: ignore[method-assign]
 
     msg = InboundMessage(
@@ -1262,7 +1313,6 @@ async def test_process_message_persists_media_only_turn_without_text(tmp_path: P
     img.write_bytes(_PNG_1X1)
 
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop._run_agent_loop = AsyncMock(side_effect=RuntimeError("boom"))  # type: ignore[method-assign]
 
     msg = InboundMessage(
@@ -1286,7 +1336,6 @@ async def test_process_message_persists_media_only_turn_without_text(tmp_path: P
 @pytest.mark.asyncio
 async def test_process_message_does_not_duplicate_early_persisted_user_message(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop._run_agent_loop = AsyncMock(return_value=_agent_run_result(
         "done",
         [
@@ -1319,7 +1368,6 @@ async def test_internal_continuation_queues_turn_without_fake_user_history(
     tmp_path: Path,
 ) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     session = loop.sessions.get_or_create("feishu:c-auto")
     session.metadata[GOAL_STATE_KEY] = {
         "status": "active",
@@ -1388,7 +1436,6 @@ async def test_internal_continuation_preserves_streaming_route_metadata(
     tmp_path: Path,
 ) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     session = loop.sessions.get_or_create("feishu:c-stream")
     session.metadata[GOAL_STATE_KEY] = {
         "status": "active",
@@ -1462,7 +1509,6 @@ async def test_websocket_internal_continuation_keeps_single_visible_run(
     tmp_path: Path,
 ) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     session = loop.sessions.get_or_create("websocket:c-auto")
     session.metadata[GOAL_STATE_KEY] = {
         "status": "active",
@@ -1526,7 +1572,6 @@ async def test_websocket_internal_continuation_keeps_single_visible_run(
 @pytest.mark.asyncio
 async def test_process_message_keeps_delivery_chat_for_thread_session(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop.context.build_messages = MagicMock(  # type: ignore[method-assign]
         return_value=[
             {"role": "system", "content": "system"},
@@ -1565,7 +1610,6 @@ async def test_process_message_uses_explicit_session_for_goal_context(
     tmp_path: Path,
 ) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     chat_session = loop.sessions.get_or_create("websocket:chat-with-goal")
     chat_session.metadata[GOAL_STATE_KEY] = {
         "status": "active",
@@ -1713,7 +1757,6 @@ async def test_request_context_uses_effective_key_for_spawn_tool(tmp_path: Path)
 @pytest.mark.asyncio
 async def test_next_turn_after_crash_closes_pending_user_turn_before_new_input(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     loop.provider.chat_with_retry = AsyncMock(return_value=MagicMock())  # unused because _run_agent_loop is stubbed
 
     session = loop.sessions.get_or_create("feishu:c3")
@@ -1762,7 +1805,6 @@ async def test_stop_preserves_runtime_checkpoint_for_next_turn(tmp_path: Path) -
     from nanobot.command.router import CommandContext
 
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     checkpoint_saved = asyncio.Event()
 
@@ -1866,7 +1908,6 @@ async def test_stop_preserves_runtime_checkpoint_for_next_turn(tmp_path: Path) -
 @pytest.mark.asyncio
 async def test_system_subagent_followup_is_persisted_before_prompt_assembly(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     session = loop.sessions.get_or_create("cli:test")
     session.add_message("user", "question")
@@ -1913,11 +1954,6 @@ async def test_system_subagent_followup_is_persisted_before_prompt_assembly(tmp_
     assert request.metadata == {"subagent_task_id": "sub-1"}
     assert request.turn_id
     record_runtime.assert_called_once_with("cli:test", runtime)
-    assert len(loop.consolidator.maybe_consolidate_by_tokens.call_args_list) == 2
-    assert all(
-        call.kwargs["runtime"] is runtime
-        for call in loop.consolidator.maybe_consolidate_by_tokens.call_args_list
-    )
     initial_messages = seen["initial_messages"]
     assert isinstance(initial_messages, list)
     non_system = [m for m in initial_messages if m.get("role") != "system"]
@@ -1952,7 +1988,6 @@ async def test_system_subagent_followup_is_persisted_before_prompt_assembly(tmp_
 @pytest.mark.asyncio
 async def test_turn_usage_is_persisted_with_the_saved_session(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
     turn_usage = LLMUsage.reported(input_tokens=64, output_tokens=9)
 
     async def fake_run_agent_loop(transcript_input, **_kwargs):
@@ -1978,9 +2013,6 @@ async def test_turn_usage_is_persisted_with_the_saved_session(tmp_path: Path) ->
 @pytest.mark.asyncio
 async def test_system_subagent_followup_does_not_log_content(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(  # type: ignore[method-assign]
-        return_value=False
-    )
 
     async def fake_run_agent_loop(transcript_input, **_kwargs):
         initial_messages = _assembled_messages(loop.context, transcript_input)
@@ -2017,9 +2049,6 @@ async def test_system_subagent_followup_does_not_log_content(tmp_path: Path) ->
 @pytest.mark.asyncio
 async def test_system_subagent_followup_uses_common_turn_lifecycle(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(  # type: ignore[method-assign]
-        return_value=False
-    )
     visited: list[str] = []
 
     for name in (
@@ -2081,7 +2110,6 @@ async def test_system_subagent_followup_uses_common_turn_lifecycle(tmp_path: Pat
 @pytest.mark.asyncio
 async def test_multiple_subagent_followups_all_persist_as_standalone_history(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     async def fake_run_agent_loop(transcript_input, **_kwargs):
         initial_messages = _assembled_messages(loop.context, transcript_input)
@@ -2207,7 +2235,6 @@ async def test_request_context_passes_thread_session_key_to_spawn(tmp_path: Path
 @pytest.mark.asyncio
 async def test_system_subagent_followup_uses_thread_session_and_slack_metadata(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     thread_session = loop.sessions.get_or_create("slack:C123:1700.42")
     thread_session.add_message("user", "thread question")
@@ -2266,7 +2293,6 @@ async def test_system_subagent_followup_uses_thread_session_and_slack_metadata(t
 @pytest.mark.asyncio
 async def test_turn_after_unanswered_user_keeps_tool_call_pairing(tmp_path: Path) -> None:
     loop = _make_full_loop(tmp_path)
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     session = loop.sessions.get_or_create("feishu:c-merge")
     session.add_message("user", "earlier question that never got an answer")

tests/agent/test_loop_session_policy.py

@@ -46,7 +46,6 @@ def _loop(tmp_path, responses: list[str], **kwargs) -> AgentLoop:
 async def test_transient_session_keeps_history_without_persisting_or_durable_tools(tmp_path) -> None:
     loop = _loop(tmp_path, ["first answer", "second answer"])
     loop.context.memory.write_memory("private durable memory")
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock()
     key = "websocket:transient-test"
     loop.sessions.get_or_create_transient(
         key,
@@ -71,7 +70,6 @@ async def test_transient_session_keeps_history_without_persisting_or_durable_too
         "assistant",
     ]
     assert loop.sessions.read_session_file(key) is None
-    loop.consolidator.maybe_consolidate_by_tokens.assert_not_awaited()
 
 
 @pytest.mark.asyncio

tests/agent/test_runner_core.py

@@ -61,7 +61,10 @@ def test_initial_transcript_is_built_from_structured_turn_input() -> None:
         max_tool_result_chars=_MAX_TOOL_RESULT_CHARS,
     )
 
-    assert AgentRunner._initial_transcript(spec) == expected
+    messages, compaction = AgentRunner._initial_transcript_and_compaction(spec)
+
+    assert messages == expected
+    assert compaction is None
     transcript_builder.assert_called_once_with(transcript_input)
 
 

tests/agent/test_runner_governance.py

@@ -7,13 +7,14 @@ from unittest.mock import AsyncMock, MagicMock, patch
 import pytest
 
 from agent.runner_helpers import make_run_spec
+from nanobot.agent.context import TranscriptInput
 from nanobot.agent.context_governance import (
     BACKFILL_CONTENT,
     ContextGovernanceConfig,
     ContextGovernor,
     ContextWindowExceededError,
 )
-from nanobot.agent.runner import AgentRunSpec
+from nanobot.agent.runner import AgentRunner, AgentRunSpec
 from nanobot.config.schema import AgentDefaults
 from nanobot.providers.base import (
     LLMProvider,
@@ -22,10 +23,23 @@ from nanobot.providers.base import (
     ProviderConversationState,
     ToolCallRequest,
 )
+from nanobot.session.summary import SUMMARY_CONTINUATION_TEXT
 
 _MAX_TOOL_RESULT_CHARS = AgentDefaults().max_tool_result_chars
 
 
+def _build_transcript(transcript: TranscriptInput) -> list[dict]:
+    system = (
+        transcript.session_summary["text"]
+        if transcript.session_summary is not None
+        else "system"
+    )
+    messages = [{"role": "system", "content": system}, *transcript.history]
+    if transcript.current_message is not None:
+        messages.append({"role": transcript.current_role, "content": transcript.current_message})
+    return messages
+
+
 def _governance_config(
     provider,
     tools,
@@ -97,13 +111,16 @@ async def test_runner_locally_fits_oversized_initial_transcript(monkeypatch):
     tools = MagicMock()
     tools.get_definitions.return_value = []
     old_content = "x" * 20_000
-    monkeypatch.setattr(
-        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
-        lambda _provider, _model, messages, _tools: (
+    estimate = MagicMock(
+        side_effect=lambda _provider, _model, messages, _tools: (
             (600, "test-counter")
             if any(message.get("content") == old_content for message in messages)
             else (100, "test-counter")
-        ),
+        )
+    )
+    monkeypatch.setattr(
+        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
+        estimate,
     )
 
     result = await AgentRunner().run(make_run_spec(
@@ -127,9 +144,441 @@ async def test_runner_locally_fits_oversized_initial_transcript(monkeypatch):
         {"role": "system", "content": "system"},
         {"role": "user", "content": "continue"},
     ]
+    estimated_messages = [call.args[2] for call in estimate.call_args_list]
+    assert sum(
+        any(message.get("content") == old_content for message in messages)
+        for messages in estimated_messages
+    ) == 1
+    assert len(estimated_messages) == 3
     assert any(message.get("content") == old_content for message in result.messages)
 
 
+async def test_runner_summarizes_history_and_preserves_current_input(monkeypatch):
+    provider = MagicMock(spec=LLMProvider)
+    provider.can_resume_conversation_state.return_value = True
+    prior_state = ProviderConversationState(
+        kind="openai_responses",
+        provider="openai:test",
+        model="test-model",
+        version=1,
+        payload={"items": [{"type": "message", "role": "assistant"}]},
+    )
+    candidate_state = ProviderConversationState(
+        kind="openai_responses",
+        provider="openai:test",
+        model="test-model",
+        version=1,
+        payload={"items": [{"type": "message", "role": "assistant", "fresh": True}]},
+    )
+    requests: list[tuple[list[dict], object]] = []
+
+    async def request(*, messages, provider_context, **_kwargs):
+        requests.append((messages, provider_context))
+        return LLMResponse(content="done", provider_state=candidate_state)
+
+    provider.chat_with_retry = request
+    tools = MagicMock()
+    tools.get_definitions.return_value = []
+    old_answer = "old answer " * 2_000
+    monkeypatch.setattr(
+        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
+        lambda _provider, _model, messages, _tools: (
+            (600, "test-counter")
+            if any(message.get("content") == old_answer for message in messages)
+            else (100, "test-counter")
+        ),
+    )
+    consolidate = AsyncMock(return_value="fresh checkpoint")
+    previous = {"text": "existing checkpoint", "last_active": "2026-08-30T00:00:00"}
+
+    result = await AgentRunner().run(make_run_spec(
+        provider,
+        initial_messages=None,
+        transcript_input=TranscriptInput(
+            history=[
+                {"role": "user", "content": "old question"},
+                {"role": "assistant", "content": old_answer},
+            ],
+            current_message="continue the current task",
+            session_summary=previous,
+        ),
+        transcript_builder=_build_transcript,
+        consolidate_history=consolidate,
+        provider_state=prior_state,
+        tools=tools,
+        model="test-model",
+        context_window_tokens=2_000,
+        context_block_limit=500,
+        max_tokens=100,
+        max_iterations=1,
+        max_tool_result_chars=_MAX_TOOL_RESULT_CHARS,
+    ))
+
+    consolidate.assert_awaited_once_with(
+        [
+            {"role": "system", "content": "existing checkpoint"},
+            {"role": "user", "content": "old question"},
+            {"role": "assistant", "content": old_answer},
+        ],
+        "existing checkpoint",
+    )
+    assert requests[0][0] == [
+        {"role": "system", "content": "fresh checkpoint"},
+        {"role": "user", "content": SUMMARY_CONTINUATION_TEXT},
+        {"role": "user", "content": "continue the current task"},
+    ]
+    assert requests[0][1].conversation_state is None
+    assert result.provider_state == candidate_state
+    assert result.summary_checkpoint is not None
+    assert result.summary_checkpoint.summary == "fresh checkpoint"
+    assert result.summary_checkpoint.transcript_boundary == 3
+    assert any(message.get("content") == old_answer for message in result.messages)
+
+
+async def test_runner_rejects_oversized_delta_without_summarizable_history(monkeypatch):
+    provider = MagicMock(spec=LLMProvider)
+    provider.chat_with_retry = AsyncMock()
+    tools = MagicMock()
+    tools.get_definitions.return_value = []
+    monkeypatch.setattr(
+        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
+        lambda *_args, **_kwargs: (600, "test-counter"),
+    )
+    consolidate = AsyncMock(return_value=None)
+
+    with pytest.raises(ContextWindowExceededError):
+        await AgentRunner().run(make_run_spec(
+            provider,
+            initial_messages=None,
+            transcript_input=TranscriptInput(
+                history=[],
+                current_message="current input is the entire oversized delta",
+            ),
+            transcript_builder=_build_transcript,
+            consolidate_history=consolidate,
+            tools=tools,
+            model="test-model",
+            context_window_tokens=2_000,
+            context_block_limit=500,
+            max_tokens=100,
+            max_iterations=1,
+            max_tool_result_chars=_MAX_TOOL_RESULT_CHARS,
+        ))
+
+    consolidate.assert_awaited_once_with(
+        [{"role": "system", "content": "system"}],
+        None,
+    )
+    provider.chat_with_retry.assert_not_awaited()
+
+
+async def test_runner_governs_history_before_summarizing_it(monkeypatch):
+    provider = MagicMock(spec=LLMProvider)
+    provider.can_resume_conversation_state.return_value = False
+    provider.chat_with_retry = AsyncMock(return_value=LLMResponse(content="done"))
+    tools = MagicMock()
+    tools.get_definitions.return_value = []
+    monkeypatch.setattr(
+        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
+        lambda _provider, _model, messages, _tools: (
+            (100, "test-counter")
+            if messages[0].get("content") == "fresh checkpoint"
+            else (600, "test-counter")
+        ),
+    )
+    consolidate = AsyncMock(return_value="fresh checkpoint")
+
+    await AgentRunner().run(make_run_spec(
+        provider,
+        initial_messages=None,
+        transcript_input=TranscriptInput(
+            history=[
+                {"role": "user", "content": "inspect"},
+                {
+                    "role": "assistant",
+                    "content": "",
+                    "tool_calls": [{
+                        "id": "call-missing",
+                        "type": "function",
+                        "function": {"name": "inspect", "arguments": "{}"},
+                    }],
+                },
+            ],
+            current_message="continue",
+        ),
+        transcript_builder=_build_transcript,
+        consolidate_history=consolidate,
+        tools=tools,
+        model="test-model",
+        context_window_tokens=2_000,
+        context_block_limit=500,
+        max_tokens=100,
+        max_iterations=1,
+        max_tool_result_chars=_MAX_TOOL_RESULT_CHARS,
+    ))
+
+    summarized = consolidate.await_args.args[0]
+    assert [message["role"] for message in summarized] == [
+        "system", "user", "assistant", "tool",
+    ]
+    assert summarized[-1]["tool_call_id"] == "call-missing"
+    assert summarized[-1]["content"] == BACKFILL_CONTENT
+
+
+@pytest.mark.parametrize(
+    ("scope", "expected_contents", "expected_boundary"),
+    [
+        ("prior_context", ["system", "accepted question", "accepted answer"], 3),
+        (
+            "current_request",
+            ["system", "accepted question", "accepted answer", "inspect the project"],
+            4,
+        ),
+    ],
+)
+async def test_native_compaction_uses_provider_request_boundary(
+    monkeypatch,
+    scope,
+    expected_contents,
+    expected_boundary,
+):
+    provider = MagicMock(spec=LLMProvider)
+    provider.can_resume_conversation_state.return_value = False
+    compacted_state = ProviderConversationState(
+        kind="openai_responses",
+        provider="openai:test",
+        model="test-model",
+        version=1,
+        payload={"items": [{"type": "compaction", "encrypted_content": "opaque"}]},
+    )
+    provider.chat_with_retry = AsyncMock(side_effect=[
+        LLMResponse(
+            content=None,
+            tool_calls=[ToolCallRequest(id="call-1", name="inspect", arguments={})],
+            provider_compaction_applied=True,
+            provider_compaction_state=compacted_state,
+            provider_compaction_scope=scope,
+        ),
+        LLMResponse(content="done"),
+    ])
+    tools = MagicMock()
+    tools.get_definitions.return_value = []
+    tools.execute = AsyncMock(return_value="complete tool result")
+    monkeypatch.setattr(
+        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
+        lambda *_args: (100, "test-counter"),
+    )
+    consolidate = AsyncMock(return_value="portable checkpoint")
+    consolidate_native = AsyncMock(return_value="portable checkpoint")
+
+    result = await AgentRunner().run(make_run_spec(
+        provider,
+        initial_messages=None,
+        transcript_input=TranscriptInput(
+            history=[
+                {"role": "user", "content": "accepted question"},
+                {"role": "assistant", "content": "accepted answer"},
+            ],
+            current_message="inspect the project",
+        ),
+        transcript_builder=_build_transcript,
+        consolidate_history=consolidate,
+        consolidate_provider_compaction=consolidate_native,
+        tools=tools,
+        model="test-model",
+        context_window_tokens=2_000,
+        context_block_limit=500,
+        max_tokens=100,
+        max_iterations=2,
+        max_tool_result_chars=_MAX_TOOL_RESULT_CHARS,
+    ))
+
+    consolidate.assert_not_awaited()
+    consolidate_native.assert_awaited_once()
+    assert consolidate_native.await_args.args[0] == compacted_state
+    assert [
+        message["content"] for message in consolidate_native.await_args.args[1]
+    ] == expected_contents
+    assert consolidate_native.await_args.args[2] is None
+    assert result.summary_checkpoint is not None
+    assert result.summary_checkpoint.transcript_boundary == expected_boundary
+    assert result.provider_compaction_applied is True
+    assert any(message.get("content") == "inspect the project" for message in result.messages)
+    assert any(message.get("content") == "complete tool result" for message in result.messages)
+
+
+async def test_runner_keeps_current_tool_exchange_outside_summary(monkeypatch):
+    provider = MagicMock(spec=LLMProvider)
+    provider.can_resume_conversation_state.return_value = False
+    responses = [
+        LLMResponse(
+            content=None,
+            tool_calls=[ToolCallRequest(id="call-1", name="inspect", arguments={})],
+        ),
+        LLMResponse(content="done"),
+    ]
+    requests: list[list[dict]] = []
+
+    async def request(*, messages, **_kwargs):
+        requests.append(messages)
+        return responses.pop(0)
+
+    provider.chat_with_retry = request
+    tools = MagicMock()
+    tools.get_definitions.return_value = []
+    full_result = "tool-result:" + ("x" * 4_000)
+    tools.execute = AsyncMock(return_value=full_result)
+
+    def estimate(_provider, _model, messages, _tools):
+        has_tool_result = any(message.get("role") == "tool" for message in messages)
+        has_old_system = any(
+            message.get("role") == "system" and message.get("content") == "system"
+            for message in messages
+        )
+        return (600 if has_tool_result and has_old_system else 100, "test-counter")
+
+    monkeypatch.setattr(
+        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
+        estimate,
+    )
+    consolidate = AsyncMock(return_value="fresh checkpoint")
+
+    result = await AgentRunner().run(make_run_spec(
+        provider,
+        initial_messages=None,
+        transcript_input=TranscriptInput(history=[], current_message="inspect the project"),
+        transcript_builder=_build_transcript,
+        consolidate_history=consolidate,
+        tools=tools,
+        model="test-model",
+        context_window_tokens=2_000,
+        context_block_limit=500,
+        max_tokens=100,
+        max_iterations=2,
+        max_tool_result_chars=_MAX_TOOL_RESULT_CHARS,
+    ))
+
+    consolidate.assert_awaited_once_with(
+        [
+            {"role": "system", "content": "system"},
+            {"role": "user", "content": "inspect the project"},
+        ],
+        None,
+    )
+    assert [message["role"] for message in requests[1]] == [
+        "system", "user", "assistant", "tool",
+    ]
+    assert requests[1][1]["content"] == SUMMARY_CONTINUATION_TEXT
+    assert requests[1][-1]["content"] == full_result
+    assert result.summary_checkpoint is not None
+    assert result.summary_checkpoint.transcript_boundary == 2
+    assert any(message.get("content") == full_result for message in result.messages)
+
+
+async def test_repeated_pressure_advances_summary_boundary(monkeypatch):
+    provider = MagicMock(spec=LLMProvider)
+    provider.can_resume_conversation_state.return_value = False
+    provider.chat_with_retry = AsyncMock(side_effect=[
+        LLMResponse(
+            content=None,
+            tool_calls=[ToolCallRequest(id="call-1", name="inspect", arguments={})],
+        ),
+        LLMResponse(
+            content=None,
+            tool_calls=[ToolCallRequest(id="call-2", name="inspect", arguments={})],
+        ),
+        LLMResponse(content="done"),
+    ])
+    tools = MagicMock()
+    tools.get_definitions.return_value = []
+    tools.execute = AsyncMock(side_effect=["result-1", "result-2"])
+
+    def estimate(_provider, _model, messages, _tools):
+        system = messages[0].get("content")
+        contents = {message.get("content") for message in messages}
+        if "result-2" in contents:
+            return (100 if system == "checkpoint-2" else 600, "test-counter")
+        if "result-1" in contents:
+            return (100 if system == "checkpoint-1" else 600, "test-counter")
+        return 100, "test-counter"
+
+    monkeypatch.setattr(
+        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
+        estimate,
+    )
+    consolidate = AsyncMock(side_effect=["checkpoint-1", "checkpoint-2"])
+
+    result = await AgentRunner().run(make_run_spec(
+        provider,
+        initial_messages=None,
+        transcript_input=TranscriptInput(history=[], current_message="inspect"),
+        transcript_builder=_build_transcript,
+        consolidate_history=consolidate,
+        tools=tools,
+        model="test-model",
+        context_window_tokens=2_000,
+        context_block_limit=500,
+        max_tokens=100,
+        max_iterations=3,
+        max_tool_result_chars=_MAX_TOOL_RESULT_CHARS,
+    ))
+
+    assert consolidate.await_count == 2
+    assert consolidate.await_args_list[0].args[1] is None
+    assert consolidate.await_args_list[1].args[1] == "checkpoint-1"
+    second_prefix = consolidate.await_args_list[1].args[0]
+    assert second_prefix[0]["content"] == "checkpoint-1"
+    assert any(message.get("content") == "result-1" for message in second_prefix)
+    assert result.final_content == "done"
+    assert result.summary_checkpoint is not None
+    assert result.summary_checkpoint.summary == "checkpoint-2"
+    assert result.summary_checkpoint.transcript_boundary == 4
+
+
+async def test_runner_refuses_checkpoint_that_cannot_fit_with_delta(monkeypatch):
+    provider = MagicMock(spec=LLMProvider)
+    provider.chat_with_retry = AsyncMock(return_value=LLMResponse(content="unexpected"))
+    tools = MagicMock()
+    tools.get_definitions.return_value = []
+    old_answer = "old answer"
+    current_input = "current input must remain intact"
+
+    def estimate(_provider, _model, messages, _tools):
+        contents = {message.get("content") for message in messages}
+        if old_answer in contents or current_input in contents:
+            return 600, "test-counter"
+        return 100, "test-counter"
+
+    monkeypatch.setattr(
+        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
+        estimate,
+    )
+    consolidate = AsyncMock(return_value="small checkpoint")
+
+    with pytest.raises(ContextWindowExceededError):
+        await AgentRunner().run(make_run_spec(
+            provider,
+            initial_messages=None,
+            transcript_input=TranscriptInput(
+                history=[{"role": "assistant", "content": old_answer}],
+                current_message=current_input,
+            ),
+            transcript_builder=_build_transcript,
+            consolidate_history=consolidate,
+            tools=tools,
+            model="test-model",
+            context_window_tokens=2_000,
+            context_block_limit=500,
+            max_tokens=100,
+            max_iterations=1,
+            max_tool_result_chars=_MAX_TOOL_RESULT_CHARS,
+        ))
+
+    summarized = consolidate.await_args.args[0]
+    assert all(message.get("content") != current_input for message in summarized)
+    provider.chat_with_retry.assert_not_awaited()
+
+
 @pytest.mark.asyncio
 async def test_runner_governs_messages_added_by_before_iteration_hook(monkeypatch):
     from nanobot.agent.hook import AgentHook, AgentHookContext
@@ -145,7 +594,7 @@ async def test_runner_governs_messages_added_by_before_iteration_hook(monkeypatc
         "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
         lambda _provider, _model, messages, _tools: (
             (2_000, "test-counter")
-            if any(message.get("content") == oversized for message in messages)
+            if any(oversized in str(message.get("content")) for message in messages)
             else (100, "test-counter")
         ),
     )
@@ -491,6 +940,39 @@ def test_changed_messages_use_local_estimate_after_reported_usage(monkeypatch):
     estimate.assert_called_once()
 
 
+def test_resumed_provider_context_avoids_full_transcript_estimate(monkeypatch):
+    provider = MagicMock(spec=LLMProvider)
+    tools = MagicMock()
+    tools.get_definitions.return_value = []
+    spec = make_run_spec(
+        provider,
+        initial_messages=[{"role": "user", "content": "pending delta"}],
+        tools=tools,
+        model="local-model",
+        context_window_tokens=2_000,
+        context_block_limit=500,
+        max_iterations=1,
+        max_tool_result_chars=_MAX_TOOL_RESULT_CHARS,
+    )
+    monkeypatch.setattr(
+        "nanobot.agent.context_governance.estimate_prompt_tokens_chain",
+        lambda *_args, **_kwargs: (_ for _ in ()).throw(
+            AssertionError("resumed provider context must be authoritative")
+        ),
+    )
+
+    pressure = ContextGovernor().request_pressure(
+        _governance_config(provider, tools, spec),
+        spec.initial_messages,
+        LLMUsage.reported(input_tokens=900, output_tokens=10),
+        usage_matches_messages=False,
+        tool_definitions=tools.get_definitions(),
+        request_context_tokens=100,
+    )
+
+    assert pressure is None
+
+
 @pytest.mark.asyncio
 async def test_runner_counts_resumed_provider_state_before_dispatch(monkeypatch):
     from nanobot.agent.runner import AgentRunner
@@ -832,7 +1314,6 @@ async def test_backfill_repairs_model_context_without_shifting_save_turn_boundar
         model="test-model",
     )
     loop.tools.get_definitions = MagicMock(return_value=[])
-    loop.consolidator.maybe_consolidate_by_tokens = AsyncMock(return_value=False)  # type: ignore[method-assign]
 
     session = loop.sessions.get_or_create("cli:test")
     session.messages = [

tests/agent/test_runner_injections.py

@@ -750,25 +750,27 @@ async def test_pending_injection_resolves_its_own_runtime_context(tmp_path):
         ),
     ]
 
-    injected = [message for message in result.messages if message.get("role") == "user"][-1]
-    assert "follow-up from the second speaker" in str(injected["content"])
+    injected = [message for message in result.messages if message.get("role") == "user"][-2:]
+    assert str(injected[0]["content"]).startswith("follow-up from the second speaker\n\n")
+    assert str(injected[1]["content"]).startswith("another follow-up\n\n")
     model_messages = provider.chat_with_retry.await_args_list[-1].kwargs["messages"]
     assert "telegram | group-1 | user-b | message-2" in str(model_messages)
     assert "Bob | topic-7" in str(model_messages)
     assert "telegram | group-1 | user-c | message-3" in str(model_messages)
     assert "Carol | topic-7" in str(model_messages)
-    assert injected["_meta"][RUNTIME_CONTEXT_MESSAGE_META]["sources"] == [
-        "identity",
-        "identity",
-    ]
+    assert all(
+        message["_meta"][RUNTIME_CONTEXT_MESSAGE_META]["sources"] == ["identity"]
+        for message in injected
+    )
 
     loop._save_turn(session, result.messages, skip=1)
-    persisted = [message for message in session.messages if message.get("role") == "user"][-1]
-    assert "telegram | group-1 | user-b | message-2" in str(persisted["content"])
-    assert "telegram | group-1 | user-c | message-3" in str(persisted["content"])
-    assert public_history_message(persisted)["content"] == (
-        "follow-up from the second speaker\n\nanother follow-up"
-    )
+    persisted = [message for message in session.messages if message.get("role") == "user"][-2:]
+    assert "telegram | group-1 | user-b | message-2" in str(persisted[0]["content"])
+    assert "telegram | group-1 | user-c | message-3" in str(persisted[1]["content"])
+    assert [public_history_message(message)["content"] for message in persisted] == [
+        "follow-up from the second speaker",
+        "another follow-up",
+    ]
 
 
 @pytest.mark.asyncio
@@ -835,8 +837,8 @@ async def test_subagent_pending_injection_is_hidden_history_and_not_merged(tmp_p
 
 
 @pytest.mark.asyncio
-async def test_runner_merges_multiple_injected_user_messages_without_losing_media():
-    """Multiple injected follow-ups should not create lossy consecutive user messages."""
+async def test_model_request_merges_injected_user_messages_without_losing_media():
+    """The model copy may merge follow-ups while the raw transcript keeps each event."""
     from nanobot.agent.runner import AgentRunner
 
     provider = MagicMock()
@@ -895,10 +897,17 @@ async def test_runner_merges_multiple_injected_user_messages_without_losing_medi
         for block in injected["content"]
         if isinstance(block, dict)
     )
+    assert [message["content"] for message in result.messages[-3:-1]] == [
+        [
+            {"type": "image_url", "image_url": {"url": "data:image/png;base64,abc"}},
+            {"type": "text", "text": "look at this"},
+        ],
+        "and answer briefly",
+    ]
 
 
-def test_runner_merge_keeps_all_recovery_followup_ids() -> None:
-    """Merged follow-ups stay acknowledged together after a later save."""
+def test_runner_append_keeps_recovery_followups_separate() -> None:
+    """Each raw follow-up keeps its own recovery identity."""
     from nanobot.agent.runner import AgentRunner
     from nanobot.session.recovery import PENDING_FOLLOWUP_ID_KEY
 
@@ -908,10 +917,12 @@ def test_runner_merge_keeps_all_recovery_followup_ids() -> None:
         [{"role": "user", "content": "second", PENDING_FOLLOWUP_ID_KEY: "two"}],
     )
 
-    assert messages[-1][PENDING_FOLLOWUP_ID_KEY] == ["one", "two"]
+    assert [message["content"] for message in messages] == ["first", "second"]
+    assert [message[PENDING_FOLLOWUP_ID_KEY] for message in messages] == ["one", "two"]
 
 
-def test_runner_merge_preserves_runtime_markers_with_media() -> None:
+def test_model_request_merge_preserves_runtime_markers_with_media() -> None:
+    from nanobot.agent.context_governance import ContextGovernor
     from nanobot.agent.runner import AgentRunner
     from nanobot.runtime_context import (
         RUNTIME_CONTEXT_HISTORY_META,
@@ -948,8 +959,9 @@ def test_runner_merge_preserves_runtime_markers_with_media() -> None:
         },
     ])
 
-    assert len(messages) == 1
-    merged = messages[0]
+    assert len(messages) == 2
+    merged = ContextGovernor._merge_adjacent_user_messages_for_model(messages)[0]
+    assert len(messages) == 2
     assert "private first" in str(merged["content"])
     assert "private second" in str(merged["content"])
     persisted = {
@@ -1681,15 +1693,17 @@ async def test_drain_injections_after_recoverable_tool_error():
 
 @pytest.mark.asyncio
 async def test_drain_injections_on_llm_error():
-    """Pending injections should be drained when the LLM returns an error finish_reason."""
+    """A follow-up after an error stays raw and reaches the next model request."""
     from nanobot.agent.runner import AgentRunner
     from nanobot.bus.events import InboundMessage
 
     provider = MagicMock()
     call_count = {"n": 0}
+    requests: list[list[dict]] = []
 
     async def chat_with_retry(*, messages, **kwargs):
         call_count["n"] += 1
+        requests.append(messages)
         if call_count["n"] == 1:
             return LLMResponse(
                 content=None,
@@ -1713,11 +1727,20 @@ async def test_drain_injections_on_llm_error():
 
     runner = AgentRunner()
     result = await runner.run(make_run_spec(provider,
-        initial_messages=[
-            {"role": "user", "content": "hello"},
-            {"role": "assistant", "content": "previous response"},
-            {"role": "user", "content": "trigger error"},
+        initial_messages=None,
+        transcript_input=TranscriptInput(
+            history=[
+                {"role": "user", "content": "hello"},
+                {"role": "assistant", "content": "previous response"},
+                {"role": "user", "content": "trigger error"},
+            ],
+            current_message=None,
+        ),
+        transcript_builder=lambda transcript: [
+            {"role": "system", "content": "system"},
+            *transcript.history,
         ],
+        consolidate_history=AsyncMock(return_value=None),
         tools=tools,
         model="test-model",
         max_iterations=5,
@@ -1727,11 +1750,15 @@ async def test_drain_injections_on_llm_error():
 
     assert result.had_injections is True
     assert result.final_content == "recovered answer"
-    injected = [
-        m for m in result.messages
-        if m.get("role") == "user" and "follow-up after LLM error" in str(m.get("content", ""))
+    assert "follow-up after LLM error" in str(requests[1])
+    assert [
+        message["content"]
+        for message in result.messages
+        if message.get("role") == "user"
+    ][-2:] == [
+        "trigger error",
+        "follow-up after LLM error",
     ]
-    assert len(injected) == 1
 
 
 @pytest.mark.asyncio

tests/agent/test_session_manager_history.py

@@ -1,10 +1,11 @@
-from nanobot.providers.base import ProviderConversationState
 from nanobot.runtime_context import (
     RUNTIME_CONTEXT_HISTORY_META,
     RuntimeContextBlock,
     append_runtime_context,
 )
+from nanobot.session.history_visibility import HIDDEN_HISTORY_META
 from nanobot.session.manager import Session, SessionManager
+from nanobot.session.summary import SUMMARY_CONTINUATION_TEXT
 
 
 def _assert_no_orphans(history: list[dict]) -> None:
@@ -136,58 +137,6 @@ def test_legitimate_tool_pairs_preserved_after_trim():
     assert history[0]["role"] == "user"
 
 
-def test_retain_recent_legal_suffix_keeps_recent_messages():
-    session = Session(key="test:trim")
-    for i in range(10):
-        session.messages.append({"role": "user", "content": f"msg{i}"})
-
-    session.retain_recent_legal_suffix(4)
-
-    assert len(session.messages) == 4
-    assert session.messages[0]["content"] == "msg6"
-    assert session.messages[-1]["content"] == "msg9"
-
-
-def test_retain_recent_legal_suffix_adjusts_last_archived():
-    session = Session(key="test:trim-cons")
-    for i in range(10):
-        session.messages.append({"role": "user", "content": f"msg{i}"})
-    session.last_archived = 7
-
-    session.retain_recent_legal_suffix(4)
-
-    assert len(session.messages) == 4
-    assert session.last_archived == 1
-
-
-def test_retain_recent_legal_suffix_zero_clears_session():
-    session = Session(key="test:trim-zero")
-    for i in range(10):
-        session.messages.append({"role": "user", "content": f"msg{i}"})
-    session.last_archived = 5
-
-    session.retain_recent_legal_suffix(0)
-
-    assert session.messages == []
-    assert session.last_archived == 0
-
-
-def test_retain_recent_legal_suffix_keeps_legal_tool_boundary():
-    session = Session(key="test:trim-tools")
-    session.messages.append({"role": "user", "content": "old"})
-    session.messages.extend(_tool_turn("old", 0))
-    session.messages.append({"role": "user", "content": "keep"})
-    session.messages.extend(_tool_turn("keep", 0))
-    session.messages.append({"role": "assistant", "content": "done"})
-
-    session.retain_recent_legal_suffix(4)
-
-    history = session.get_history(max_messages=500)
-    _assert_no_orphans(history)
-    assert history[0]["role"] == "user"
-    assert history[0]["content"] == "keep"
-
-
 # --- last_archived > 0 ---
 
 def test_orphan_trim_with_last_archived():
@@ -635,6 +584,40 @@ def test_fork_session_allows_index_equal_to_user_count(tmp_path):
     assert [m["content"] for m in forked.messages] == ["round1", "answer1"]
 
 
+def test_fork_session_user_index_ignores_hidden_checkpoint_anchor(tmp_path):
+    manager = SessionManager(tmp_path)
+    source = manager.get_or_create("websocket:source")
+    source.add_message("user", "round1")
+    source.add_message("assistant", "answer1")
+    source.add_message("user", "round2")
+    source.add_message(
+        "user",
+        SUMMARY_CONTINUATION_TEXT,
+        **{HIDDEN_HISTORY_META: True},
+    )
+    source.add_message("assistant", "answer2")
+    source.last_archived = 3
+    source.metadata["_last_summary"] = {"text": "round1 and round2"}
+    manager.save(source)
+
+    forked = manager.fork_session_before_user_index(
+        "websocket:source",
+        "websocket:fork",
+        2,
+    )
+
+    assert forked is not None
+    assert [message["content"] for message in forked.messages] == [
+        "round1",
+        "answer1",
+        "round2",
+        SUMMARY_CONTINUATION_TEXT,
+        "answer2",
+    ]
+    assert forked.last_archived == 3
+    assert forked.metadata["_last_summary"]["text"] == "round1 and round2"
+
+
 def test_fork_session_drops_summary_when_fork_point_is_inside_archived_prefix(tmp_path):
     manager = SessionManager(tmp_path)
     source = manager.get_or_create("websocket:source")
@@ -756,44 +739,6 @@ def test_get_history_recovers_user_when_token_slice_would_be_assistant_only(monk
     assert [m["content"] for m in history] == ["u2", "a2"]
 
 
-def test_retain_recent_legal_suffix_hard_cap_with_long_non_user_chain():
-    session = Session(key="test:hard-cap-chain")
-    session.messages.append({"role": "user", "content": "u0"})
-    session.messages.append(
-        {
-            "role": "assistant",
-            "content": None,
-            "tool_calls": [
-                {"id": "c1", "type": "function", "function": {"name": "x", "arguments": "{}"}}
-            ],
-        }
-    )
-    for i in range(12):
-        session.messages.append({"role": "assistant", "content": f"a{i}"})
-
-    session.retain_recent_legal_suffix(6)
-
-    assert len(session.messages) <= 6
-
-
-def test_retain_recent_legal_suffix_can_extend_to_user_for_long_recent_turn():
-    session = Session(key="test:extend-to-user")
-    session.messages.append({"role": "user", "content": "old"})
-    session.messages.append({"role": "assistant", "content": "old answer"})
-    session.messages.append({"role": "user", "content": "record this"})
-    for i in range(4):
-        session.messages.extend(_tool_turn("recent", i))
-    session.messages.append({"role": "assistant", "content": "done"})
-
-    session.retain_recent_legal_suffix(8, extend_to_user=True)
-
-    assert len(session.messages) > 8
-    assert session.messages[0]["content"] == "record this"
-    assert session.messages[-1]["content"] == "done"
-    history = session.get_history(max_messages=500)
-    _assert_no_orphans(history)
-
-
 def test_get_history_can_extend_to_user_for_long_recent_turn():
     session = Session(key="test:history-extend-to-user")
     session.messages.append({"role": "user", "content": "old"})
@@ -828,82 +773,3 @@ def test_get_history_extend_to_user_keeps_newer_user_inside_window():
 
     assert [m["content"] for m in history] == ["new question", "new answer"]
     _assert_no_orphans(history)
-
-
-def test_retain_recent_legal_suffix_returns_dropped_messages():
-    """retain_recent_legal_suffix returns the actually-dropped messages."""
-    session = Session(
-        key="test:return-dropped",
-        provider_state=ProviderConversationState(
-            kind="openai_responses",
-            provider="openai:test",
-            model="test-model",
-            version=1,
-            payload={"items": []},
-        ),
-    )
-    for i in range(10):
-        session.messages.append({"role": "user", "content": f"msg{i}"})
-
-    result = session.retain_recent_legal_suffix(4)
-
-    assert len(result.dropped) == 6
-    assert [m["content"] for m in result.dropped] == [f"msg{i}" for i in range(6)]
-    assert len(session.messages) == 4
-    assert result.already_consolidated_count == 0
-    assert session.provider_state is None
-
-
-def test_retain_recent_legal_suffix_returns_empty_when_no_drop():
-    """No messages dropped → empty list returned."""
-    state = ProviderConversationState(
-        kind="openai_responses",
-        provider="openai:test",
-        model="test-model",
-        version=1,
-        payload={"items": []},
-    )
-    session = Session(key="test:no-drop", provider_state=state)
-    for i in range(3):
-        session.messages.append({"role": "user", "content": f"msg{i}"})
-
-    result = session.retain_recent_legal_suffix(4)
-
-    assert result.dropped == []
-    assert result.already_consolidated_count == 0
-    assert len(session.messages) == 3
-    assert session.provider_state is state
-
-
-def test_retain_recent_legal_suffix_returns_all_on_zero():
-    """max_messages=0 clears session and returns all messages."""
-    session = Session(key="test:zero-return")
-    for i in range(5):
-        session.messages.append({"role": "user", "content": f"msg{i}"})
-    session.last_archived = 3
-
-    result = session.retain_recent_legal_suffix(0)
-
-    assert len(result.dropped) == 5
-    assert result.already_consolidated_count == 3
-    assert session.messages == []
-
-
-def test_retain_recent_legal_suffix_last_archived_correct_in_else_branch():
-    """last_archived should count retained messages from the old archived prefix."""
-    session = Session(key="test:else-lc-correct")
-    # 20 messages: u0..u9, a0..a9
-    for i in range(10):
-        session.messages.append({"role": "user", "content": f"u{i}"})
-    for i in range(10):
-        session.messages.append({"role": "assistant", "content": f"a{i}"})
-    session.last_archived = 12  # u0..u9, a0, a1 archived
-
-    result = session.retain_recent_legal_suffix(4)
-
-    # Retained messages start from latest user (u9) + max_messages forward
-    # so retained = [u9, a0..a9][:4] → but these are from original indices 9..12
-    # Of those, indices 9,10,11 are < 12 (before_lc), so new_lc = 3
-    assert session.last_archived == 3
-    # already_cons should count dropped messages with original index < 12
-    assert result.already_consolidated_count == 9

tests/agent/test_session_retention.py

@@ -1,225 +0,0 @@
-from nanobot.session.manager import Session
-
-
-def _assert_no_orphans(history: list[dict]) -> None:
-    declared = {
-        tc["id"]
-        for m in history
-        if m.get("role") == "assistant"
-        for tc in (m.get("tool_calls") or [])
-    }
-    orphans = [
-        m.get("tool_call_id")
-        for m in history
-        if m.get("role") == "tool" and m.get("tool_call_id") not in declared
-    ]
-    assert orphans == [], f"orphan tool_call_ids: {orphans}"
-
-
-def _delivery(content: str) -> dict:
-    return {"role": "assistant", "content": content, "_channel_delivery": True}
-
-
-def _tool_turn(prefix: str, idx: int) -> list[dict]:
-    return [
-        {
-            "role": "assistant",
-            "content": None,
-            "tool_calls": [
-                {
-                    "id": f"{prefix}_{idx}_a",
-                    "type": "function",
-                    "function": {"name": "x", "arguments": "{}"},
-                },
-                {
-                    "id": f"{prefix}_{idx}_b",
-                    "type": "function",
-                    "function": {"name": "y", "arguments": "{}"},
-                },
-            ],
-        },
-        {"role": "tool", "tool_call_id": f"{prefix}_{idx}_a", "name": "x", "content": "ok"},
-        {"role": "tool", "tool_call_id": f"{prefix}_{idx}_b", "name": "y", "content": "ok"},
-    ]
-
-
-def _contents(messages: list[dict]) -> list[str]:
-    return [m.get("content") for m in messages]
-
-
-def _has_delivery(messages: list[dict]) -> bool:
-    return any(m.get("_channel_delivery") for m in messages)
-
-
-# --- Hard-cap trimming must preserve a proactive delivery the user replied to ---
-
-
-def test_retain_hard_cap_keeps_delivery_before_user():
-    session = Session(key="test:cap-delivery")
-    session.messages.append({"role": "user", "content": "setup"})
-    session.messages.append(_delivery("Remember to drink water"))
-    session.messages.append({"role": "user", "content": "ok"})
-    session.messages.append({"role": "assistant", "content": "great"})
-
-    session.retain_recent_legal_suffix(3)
-
-    assert _has_delivery(session.messages), "delivery dropped by hard-cap trim"
-    assert _contents(session.messages) == [
-        "Remember to drink water",
-        "ok",
-        "great",
-    ]
-
-
-def test_retain_hard_cap_matches_get_history_boundary():
-    """The trimmed suffix must start on the same message as get_history()."""
-    session = Session(key="test:cap-boundary")
-    session.messages.append({"role": "user", "content": "setup"})
-    session.messages.append(_delivery("You have 3 pending tasks"))
-    session.messages.append({"role": "user", "content": "show them"})
-    session.messages.append({"role": "assistant", "content": "done"})
-
-    expected = session.get_history(max_messages=3)
-
-    session.retain_recent_legal_suffix(3)
-
-    assert _contents(session.messages) == _contents(expected)
-
-
-def test_retain_extend_to_user_keeps_delivery_before_recovered_user():
-    session = Session(key="test:extend-delivery")
-    session.messages.append({"role": "user", "content": "setup"})
-    session.messages.append({"role": "assistant", "content": "work"})
-    session.messages.append(_delivery("Reminder: deploy at 17:00"))
-    session.messages.append({"role": "user", "content": "ok"})
-    session.messages.append({"role": "assistant", "content": "a1"})
-    session.messages.append({"role": "assistant", "content": "a2"})
-    session.messages.append({"role": "assistant", "content": "a3"})
-
-    session.retain_recent_legal_suffix(3, extend_to_user=True)
-
-    assert _has_delivery(session.messages), "delivery dropped by extend_to_user trim"
-    assert session.messages[0]["content"] == "Reminder: deploy at 17:00"
-    assert session.messages[-1]["content"] == "a3"
-
-
-def test_retain_extend_to_user_matches_get_history_boundary():
-    session = Session(key="test:extend-boundary")
-    session.messages.append({"role": "user", "content": "setup"})
-    session.messages.append({"role": "assistant", "content": "work"})
-    session.messages.append(_delivery("Reminder: review the draft"))
-    session.messages.append({"role": "user", "content": "ok"})
-    session.messages.append({"role": "assistant", "content": "a1"})
-    session.messages.append({"role": "assistant", "content": "a2"})
-    session.messages.append({"role": "assistant", "content": "a3"})
-
-    expected = session.get_history(max_messages=3, extend_to_user=True)
-
-    session.retain_recent_legal_suffix(3, extend_to_user=True)
-
-    assert _contents(session.messages) == _contents(expected)
-
-
-def test_retain_extend_to_user_does_not_extend_delivery_only_tail():
-    session = Session(key="test:extend-no-user")
-    for i in range(4):
-        session.messages.append(_delivery(f"notification {i}"))
-
-    session.retain_recent_legal_suffix(3, extend_to_user=True)
-
-    assert _contents(session.messages) == [
-        "notification 1",
-        "notification 2",
-        "notification 3",
-    ]
-
-
-# --- Only the immediately-preceding delivery is part of the anchor ---
-
-
-def test_retain_keeps_only_immediate_delivery():
-    session = Session(key="test:multi-delivery")
-    session.messages.append({"role": "user", "content": "setup"})
-    session.messages.append(_delivery("old scheduled note"))
-    session.messages.append(_delivery("new scheduled note"))
-    session.messages.append({"role": "user", "content": "ok"})
-    session.messages.append({"role": "assistant", "content": "great"})
-
-    session.retain_recent_legal_suffix(3)
-
-    kept = _contents(session.messages)
-    assert kept == ["new scheduled note", "ok", "great"], kept
-
-
-def test_retain_drops_delivery_not_adjacent_to_anchor_user():
-    """A delivery that does not immediately precede the retained user turn is
-    not part of the anchor and should not be force-retained."""
-    session = Session(key="test:nonadjacent")
-    session.messages.append({"role": "user", "content": "setup"})
-    session.messages.append(_delivery("unrelated scheduled note"))
-    session.messages.append({"role": "assistant", "content": "reply"})
-    session.messages.append({"role": "user", "content": "ok"})
-    session.messages.append({"role": "assistant", "content": "great"})
-
-    session.retain_recent_legal_suffix(2)
-
-    assert not _has_delivery(session.messages)
-    assert _contents(session.messages) == ["ok", "great"]
-
-
-def test_compact_probe_keeps_delivery_in_visible_suffix():
-    """compact_idle_session() trims a probe copy with extend_to_user=True; the
-    visible suffix it keeps must still contain the delivery message."""
-    tail = [
-        {"role": "user", "content": "setup"},
-        {"role": "assistant", "content": "work"},
-        _delivery("Reminder: deploy at 17:00"),
-        {"role": "user", "content": "ok"},
-        {"role": "assistant", "content": "a1"},
-        {"role": "assistant", "content": "a2"},
-        {"role": "assistant", "content": "a3"},
-    ]
-    probe = Session(key="test:probe", messages=tail)
-
-    probe.retain_recent_legal_suffix(3, extend_to_user=True)
-
-    assert _has_delivery(probe.messages)
-    assert probe.messages[0]["content"] == "Reminder: deploy at 17:00"
-
-
-# --- Trimming must stay coherent with the rest of replay ---
-
-
-def test_retain_then_replay_keeps_delivery_and_no_orphans():
-    session = Session(key="test:replay-after-trim")
-    session.messages.append({"role": "user", "content": "setup"})
-    session.messages.append(_delivery("You have 3 pending tasks"))
-    session.messages.append({"role": "user", "content": "show them"})
-    session.messages.extend(_tool_turn("cur", 0))
-    session.messages.append({"role": "assistant", "content": "done"})
-
-    session.retain_recent_legal_suffix(6)
-
-    assert _has_delivery(session.messages)
-    history = session.get_history(max_messages=500)
-    _assert_no_orphans(history)
-    assert any(m.get("content") == "You have 3 pending tasks" for m in history)
-
-
-def test_retain_keeps_delivery_when_user_inside_window():
-    """When the capped window already contains a user, its immediately
-    preceding delivery must stay attached to it."""
-    session = Session(key="test:window-user")
-    session.messages.append({"role": "user", "content": "setup"})
-    session.messages.append({"role": "assistant", "content": "a0"})
-    session.messages.append(_delivery("Reminder"))
-    session.messages.append({"role": "user", "content": "ok"})
-    session.messages.append({"role": "assistant", "content": "a1"})
-    session.messages.append({"role": "assistant", "content": "a2"})
-
-    expected = session.get_history(max_messages=4)
-
-    session.retain_recent_legal_suffix(4)
-
-    assert _has_delivery(session.messages)
-    assert _contents(session.messages) == _contents(expected)

tests/agent/test_unified_session.py

@@ -28,7 +28,7 @@ from nanobot.command.router import CommandContext, CommandRouter
 from nanobot.config.schema import AgentDefaults, Config
 from nanobot.providers.base import GenerationSettings
 from nanobot.session.keys import UNIFIED_SESSION_KEY
-from nanobot.session.manager import Session, SessionManager
+from nanobot.session.manager import SessionManager
 from nanobot.utils.llm_runtime import LLMRuntime
 
 # ---------------------------------------------------------------------------
@@ -334,116 +334,6 @@ class TestCmdNewUnifiedSession:
         assert len(sessions.get_or_create("discord:999").messages) == 1
 
 
-# ---------------------------------------------------------------------------
-# TestConsolidationUnaffectedByUnifiedSession — consolidation is key-agnostic
-# ---------------------------------------------------------------------------
-
-class TestConsolidationUnaffectedByUnifiedSession:
-    """maybe_consolidate_by_tokens() behaviour is identical regardless of session key."""
-
-    @pytest.mark.asyncio
-    async def test_consolidation_skips_empty_session_for_unified_key(self):
-        """Empty unified:default session → consolidation exits immediately, archive not called."""
-        from nanobot.agent.memory import Consolidator, MemoryStore
-
-        store = MagicMock(spec=MemoryStore)
-        mock_provider = MagicMock()
-        mock_provider.chat_with_retry = AsyncMock(return_value=MagicMock(content="summary"))
-        runtime = _runtime(mock_provider)
-        # Use spec= so MagicMock doesn't auto-generate AsyncMock for non-async methods,
-        # which would leave unawaited coroutines and trigger RuntimeWarning.
-        sessions = MagicMock(spec=SessionManager)
-
-        consolidator = Consolidator(
-            store=store,
-            sessions=sessions,
-            build_messages=MagicMock(return_value=[]),
-            get_tool_definitions=MagicMock(return_value=[]),
-        )
-        consolidator.archive_session = AsyncMock()
-
-        session = Session(key="unified:default")
-        session.messages = []
-        sessions.get_or_create.return_value = session
-
-        await consolidator.maybe_consolidate_by_tokens(session, runtime=runtime)
-
-        consolidator.archive_session.assert_not_called()
-
-    @pytest.mark.asyncio
-    async def test_consolidation_behaviour_identical_for_any_key(self):
-        """Archive call count is the same for 'telegram:123' and 'unified:default'
-        under identical token conditions."""
-        from nanobot.agent.memory import Consolidator, MemoryStore
-
-        archive_calls: dict[str, int] = {}
-
-        for key in ("telegram:123", "unified:default"):
-            store = MagicMock(spec=MemoryStore)
-            mock_provider = MagicMock()
-            mock_provider.chat_with_retry = AsyncMock(return_value=MagicMock(content="summary"))
-            runtime = _runtime(mock_provider)
-            sessions = MagicMock(spec=SessionManager)
-
-            consolidator = Consolidator(
-                store=store,
-                sessions=sessions,
-                build_messages=MagicMock(return_value=[]),
-                get_tool_definitions=MagicMock(return_value=[]),
-            )
-
-            session = Session(key=key)
-            session.messages = []  # empty → exits immediately for both keys
-            sessions.get_or_create.return_value = session
-
-            consolidator.archive_session = AsyncMock()
-            await consolidator.maybe_consolidate_by_tokens(
-                session,
-                runtime=runtime,
-            )
-            archive_calls[key] = consolidator.archive_session.call_count
-
-        assert archive_calls["telegram:123"] == archive_calls["unified:default"] == 0
-
-    @pytest.mark.asyncio
-    async def test_consolidation_triggers_when_over_budget_unified_key(self):
-        """When tokens exceed budget, consolidation attempts to find a boundary —
-        behaviour is identical to any other session key."""
-        from nanobot.agent.memory import Consolidator, MemoryStore
-
-        store = MagicMock(spec=MemoryStore)
-        mock_provider = MagicMock()
-        runtime = _runtime(mock_provider)
-        sessions = MagicMock(spec=SessionManager)
-
-        consolidator = Consolidator(
-            store=store,
-            sessions=sessions,
-            build_messages=MagicMock(return_value=[]),
-            get_tool_definitions=MagicMock(return_value=[]),
-        )
-
-        session = Session(key="unified:default")
-        session.messages = [{"role": "user", "content": "msg"}]
-        sessions.get_or_create.return_value = session
-
-        # Simulate over-budget: estimated > budget
-        consolidator.estimate_session_prompt_tokens = MagicMock(return_value=(950, "tiktoken"))
-        # No valid boundary found → returns gracefully without archiving
-        consolidator.pick_consolidation_boundary = MagicMock(return_value=None)
-        consolidator.archive_session = AsyncMock()
-
-        await consolidator.maybe_consolidate_by_tokens(session, runtime=runtime)
-
-        # estimate was called (consolidation was attempted)
-        consolidator.estimate_session_prompt_tokens.assert_called_once_with(
-            session,
-            runtime=runtime,
-        )
-        # but archive was not called (no valid boundary)
-        consolidator.archive_session.assert_not_called()
-
-
 # ---------------------------------------------------------------------------
 # TestStopCommandWithUnifiedSession — /stop command integration
 # ---------------------------------------------------------------------------

tests/cli/test_commands.py

@@ -1837,12 +1837,6 @@ def test_agent_workspace_override_wins_over_config_workspace(mock_agent_runtime,
     assert passed_config.workspace_path == workspace_path
 
 
-def test_heartbeat_retains_recent_messages_by_default():
-    config = Config()
-
-    assert config.gateway.heartbeat.keep_recent_messages == 8
-
-
 @pytest.mark.parametrize(
     "content, expected",
     [
@@ -2101,7 +2095,7 @@ def _patch_cli_command_runtime(
         monkeypatch.setattr("nanobot.config.paths.get_cron_dir", get_cron_dir)
 
 
-def test_heartbeat_empty_response_still_retains_recent_messages(
+def test_heartbeat_empty_response_is_not_evaluated(
     monkeypatch, tmp_path: Path,
 ) -> None:
     config_file = _write_instance_config(tmp_path)
@@ -2119,21 +2113,9 @@ def test_heartbeat_empty_response_still_retains_recent_messages(
     bus.publish_outbound = AsyncMock()
     seen: dict[str, object] = {}
 
-    class _FakeSession:
-        def retain_recent_legal_suffix(self, limit: int) -> None:
-            seen["retained_limit"] = limit
-
     class _FakeSessionManager:
         def __init__(self, _workspace: Path) -> None:
-            self.session = _FakeSession()
-            seen["heartbeat_session"] = self.session
-
-        def get_or_create(self, key: str) -> _FakeSession:
-            seen["session_key"] = key
-            return self.session
-
-        def save(self, session: _FakeSession) -> None:
-            seen["saved_session"] = session
+            pass
 
         def list_sessions(self) -> list[dict[str, str]]:
             return [{"key": "telegram:u1"}]
@@ -2199,9 +2181,6 @@ def test_heartbeat_empty_response_still_retains_recent_messages(
     response = asyncio.run(cron.on_job(CronJob(id="heartbeat", name="heartbeat")))
 
     assert response is None
-    assert seen["session_key"] == "heartbeat"
-    assert seen["retained_limit"] == config.gateway.heartbeat.keep_recent_messages
-    assert seen["saved_session"] is seen["heartbeat_session"]
 
 
 def test_webui_yes_creates_config_and_enables_local_websocket(

tests/config/test_gateway_config.py

@@ -13,3 +13,12 @@ def test_gateway_restart_mode_accepts_camel_alias():
 def test_gateway_restart_mode_rejects_unknown_value():
     with pytest.raises(ValueError):
         GatewayConfig(restart_mode="service")
+
+
+def test_heartbeat_ignores_removed_retention_limit():
+    config = Config.model_validate(
+        {"gateway": {"heartbeat": {"keepRecentMessages": 8}}}
+    )
+
+    heartbeat = config.model_dump(by_alias=True)["gateway"]["heartbeat"]
+    assert "keepRecentMessages" not in heartbeat

tests/providers/test_openai_codex_provider.py

@@ -22,7 +22,10 @@ from nanobot.providers.openai_codex_provider import (
     _request_codex,
     _should_retry_status,
 )
-from nanobot.providers.openai_responses import build_responses_state
+from nanobot.providers.openai_responses import (
+    build_responses_state,
+    responses_state_items,
+)
 from nanobot.providers.registry import find_by_name
 
 
@@ -811,12 +814,33 @@ async def test_codex_compacts_state_at_ninety_percent_before_next_request(
     )
 
     assert response.content == "done"
-    assert len(bodies) == 2
-    assert bodies[0]["input"][-1] == {"type": "compaction_trigger"}
-    assert bodies[1]["input"][-1] == {
+    assert response.provider_compaction_applied is True
+    assert response.provider_compaction_state is not None
+    assert response.provider_compaction_scope == "prior_context"
+    assert responses_state_items(response.provider_compaction_state) == [{
         "type": "compaction",
         "encrypted_content": "compacted opaque state",
-    }
+    }]
+    assert len(bodies) == 2
+    assert bodies[0]["input"][-1] == {"type": "compaction_trigger"}
+    assert not any(
+        item.get("role") == "user"
+        and "new question" in str(item.get("content"))
+        for item in bodies[0]["input"]
+    )
+    assert {
+        "type": "compaction",
+        "encrypted_content": "compacted opaque state",
+    } in bodies[1]["input"]
+    assert bodies[1]["input"].index({
+        "type": "compaction",
+        "encrypted_content": "compacted opaque state",
+    }) < next(
+        index
+        for index, item in enumerate(bodies[1]["input"])
+        if item.get("role") == "user"
+        and "new question" in str(item.get("content"))
+    )
     assert not any(
         item.get("type") == "reasoning"
         for item in bodies[1]["input"]

tests/providers/test_openai_responses.py

@@ -712,6 +712,45 @@ class TestParseResponseOutput:
         assert result.provider_state is not None
         assert responses_state_items(result.provider_state) == [*input_items, *output]
 
+    def test_marks_only_a_new_response_compaction(self):
+        compacted = parse_response_output(
+            {
+                "output": [
+                    {"type": "compaction", "encrypted_content": "opaque"},
+                    {"type": "message", "role": "assistant", "content": "done"},
+                ],
+                "status": "completed",
+                "usage": {},
+            },
+            state_provider="openai:test",
+            state_model="gpt-5.6",
+            state_input_items=[{"role": "user", "content": "old"}],
+        )
+        replayed = parse_response_output(
+            {
+                "output": [
+                    {"type": "message", "role": "assistant", "content": "continued"},
+                ],
+                "status": "completed",
+                "usage": {},
+            },
+            state_provider="openai:test",
+            state_model="gpt-5.6",
+            state_input_items=[
+                {"type": "compaction", "encrypted_content": "opaque"},
+            ],
+        )
+
+        assert compacted.provider_compaction_applied is True
+        assert compacted.provider_compaction_state is not None
+        assert compacted.provider_compaction_scope == "current_request"
+        assert responses_state_items(compacted.provider_compaction_state) == [
+            {"type": "compaction", "encrypted_content": "opaque"},
+        ]
+        assert replayed.provider_compaction_applied is False
+        assert replayed.provider_compaction_state is None
+        assert replayed.provider_compaction_scope is None
+
 
 class TestResponsesConversationState:
     def test_server_compaction_prunes_superseded_prefix(self):

tests/test_nanobot_facade.py

@@ -1412,7 +1412,6 @@ async def test_sessions_ingest_imports_transcript_without_running_model(tmp_path
     config_path = _write_config(tmp_path)
     bot = Nanobot.from_config(config_path, workspace=tmp_path)
     bot._loop.process_direct = AsyncMock()
-    bot._loop.consolidator.maybe_consolidate_by_tokens = AsyncMock()
 
     snapshot = await bot.sessions.ingest(
         "sdk:history",
@@ -1442,7 +1441,6 @@ async def test_sessions_ingest_imports_transcript_without_running_model(tmp_path
     assert snapshot.messages[0]["source"] == "longmemeval"
     assert snapshot.messages[1]["source"] == "longmemeval"
     bot._loop.process_direct.assert_not_called()
-    bot._loop.consolidator.maybe_consolidate_by_tokens.assert_not_called()
 
     reloaded = bot.sessions.get("sdk:history")
     assert reloaded is not None
@@ -1635,12 +1633,13 @@ async def test_runtime_helpers_expose_model_workspace_and_compact(tmp_path):
     runtime = bot._loop.llm_runtime()
     bot._loop.runtime_for_session = MagicMock(return_value=runtime)  # type: ignore[method-assign]
 
-    bot._loop.consolidator.maybe_consolidate_by_tokens = AsyncMock()
+    compact_session = AsyncMock()
+    bot._loop.consolidator.compact_idle_session = compact_session
     snapshot = await bot.runtime.compact_session("sdk:history")
     assert snapshot.key == "sdk:history"
-    assert (
-        bot._loop.consolidator.maybe_consolidate_by_tokens.await_args.kwargs["runtime"]
-        is runtime
+    compact_session.assert_awaited_once_with(
+        "sdk:history",
+        runtime=runtime,
     )
     assert bot.runtime.model == bot._loop.model
     assert bot.runtime.workspace == tmp_path

webui/src/lib/types.ts

@@ -714,7 +714,6 @@ export interface SettingsPayload {
     heartbeat: {
       enabled: boolean;
       interval_s: number;
-      keep_recent_messages: number;
     };
     dream: {
       schedule: string;

webui/src/tests/app-layout.test.tsx

@@ -131,7 +131,6 @@ function baseSettingsPayload() {
       heartbeat: {
         enabled: true,
         interval_s: 1800,
-        keep_recent_messages: 8,
       },
       dream: {
         schedule: "every 2h",
@@ -2470,7 +2469,6 @@ describe("App layout", () => {
                 heartbeat: {
                   enabled: true,
                   interval_s: 1800,
-                  keep_recent_messages: 8,
                 },
                 dream: {
                   schedule: "every 2h",
@@ -2960,7 +2958,6 @@ describe("App layout", () => {
                 heartbeat: {
                   enabled: true,
                   interval_s: 1800,
-                  keep_recent_messages: 8,
                 },
                 dream: {
                   schedule: "every 2h",

webui/src/tests/settings-test-utils.tsx

@@ -91,7 +91,6 @@ export function settingsPayload(): SettingsPayload {
       heartbeat: {
         enabled: true,
         interval_s: 1800,
-        keep_recent_messages: 8,
       },
       dream: {
         schedule: "every 2h",

webui/src/tests/thread-shell.test.tsx

@@ -368,7 +368,6 @@ function modelSettings(model: string, provider: string): SettingsPayload {
       heartbeat: {
         enabled: true,
         interval_s: 1800,
-        keep_recent_messages: 8,
       },
       dream: {
         schedule: "every 2h",