12 Commits

Author SHA1 Message Date
chengyongru
da08dee144 feat(provider): show cache hit rate in /status (#2645) 2026-04-02 12:51:45 +08:00
Xubin Ren
f450c6ef6c fix(channel): preserve threaded streaming context 2026-03-31 19:26:07 +08:00
Xubin Ren
351e3720b6 test(agent): cover disabled subagent exec tool
Add a regression test for the maintainer fix so subagents cannot register ExecTool when exec support is disabled.

Made-with: Cursor
2026-03-31 12:14:28 +08:00
Xubin Ren
758c4e74c9 fix(agent): preserve LoopHook error semantics when extra hooks are present 2026-03-31 02:19:29 +08:00
sontianye
f08de72f18 feat(agent): add CompositeHook for composable lifecycle hooks
Introduce a CompositeHook that fans out lifecycle callbacks to an
ordered list of AgentHook instances with per-hook error isolation.
Extract the nested _LoopHook and _SubagentHook to module scope as
public LoopHook / SubagentHook so downstream users can subclass or
compose them.  Add `hooks` parameter to AgentLoop.__init__ for
registering custom hooks at construction time.

Closes #2603
2026-03-31 02:19:29 +08:00
Xubin Ren
5bf0f6fe7d refactor: unify agent runner lifecycle hooks 2026-03-27 12:41:17 +08:00
Xubin Ren
e7d371ec1e refactor: extract shared agent runner and preserve subagent progress on failure 2026-03-27 02:49:43 +08:00
Xubin Ren
3f71014b7c fix(agent): use configured timezone when registering cron tool
Read the default timezone from the agent context when wiring the cron tool so startup no longer depends on an out-of-scope local variable. Add a regression test to ensure AgentLoop passes the configured timezone through to cron.

Made-with: Cursor
2026-03-25 22:07:14 +08:00
Xubin Ren
b5302b6f3d refactor(provider): preserve extra_content verbatim for Gemini thought_signature round-trip
Replace the flatten/unflatten approach (merging extra_content.google.*
into provider_specific_fields then reconstructing) with direct pass-through:
parse extra_content as-is, store on ToolCallRequest.extra_content, serialize
back untouched.  This is lossless, requires no hardcoded field names, and
covers all three parsing branches (str, dict, SDK object) plus streaming.
2026-03-25 10:00:29 +08:00
Yohei Nishikubo
af84b1b8c0 fix(Gemini): update ToolCallRequest and OpenAICompatProvider to handle thought signatures in extra_content 2026-03-25 10:00:29 +08:00
Xubin Ren
3dfdab704e refactor: replace litellm with native openai + anthropic SDKs
- Remove litellm dependency entirely (supply chain risk mitigation)
- Add AnthropicProvider (native SDK) and OpenAICompatProvider (unified)
- Merge CustomProvider into OpenAICompatProvider, delete custom_provider.py
- Add ProviderSpec.backend field for declarative provider routing
- Remove _resolve_model, find_gateway, find_by_model (dead heuristics)
- Pass resolved spec directly into provider — zero internal lookups
- Stub out litellm-dependent model database (cli/models.py)
- Add anthropic>=0.45.0 to dependencies, remove litellm
- 593 tests passed, net -1034 lines
2026-03-25 01:58:48 +08:00
chengyongru
72acba5d27 refactor(tests): optimize unit test structure 2026-03-24 15:12:22 +08:00