feat(skill-review): strict-mode prompt + raise min_tool_calls 3 to 5
The review agent was creating too many low-value skills. Three changes tighten the funnel: - review_worker prompt: rewritten in "保守模式". Default decision is do-nothing. To create, all five AND conditions must hold (>=5 tool calls with step dependencies, an explicit fail->fix->success episode, cross-scenario reusability, no >=70% overlap with skill_list, can produce >=300 chars of substance). Adds 7 explicit NO categories (single-tool loops, one-off queries, LLM common knowledge, etc.) and a 4-question self-check before any create. Hard cap of 1 new skill per review, with patch preferred over create. - min_tool_calls: 3 -> 5. Matches the prompt's ">=5 steps" hard condition so trivial turns never reach the LLM reviewer. - Updated config.yaml, the default-value assertion in test_skill_review_worker, and the middleware docstring's "(default 3)" annotation to keep all three references in sync. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
1 parent
539a8bcec8
commit
c493639119
5 files changed
+114
-77
No files matched your search
+1
-1
@@ -5,7 +5,7 @@ class ReviewWorkerConfig(BaseModel):
|
||||
"""Configuration for the post-turn background skill review worker."""
|
||||
|
||||
enabled: bool = Field(default=False, description="Whether the review worker is active.")
|
||||
min_tool_calls: int = Field(default=3, description="Minimum tool calls in a turn to trigger review.")
|
||||
min_tool_calls: int = Field(default=5, description="Minimum tool calls in a turn to trigger review. Matches the prompt's '≥5 steps' hard condition so simple queries never reach the LLM reviewer.")
|
||||
max_messages: int = Field(default=60, description="Max messages passed to the review LLM.")
|
||||
model_name: str | None = Field(default=None, description="Model to use for review; defaults to primary model.")
|
||||
timeout_seconds: int = Field(default=120, description="Overall timeout for a single review run.")
|
||||
|
||||
Reference in new issue
Block a user