Beaver/beaver_project
6
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

feat(skill-learning): preserve base skill during synthesis

Browse Source
This commit is contained in:
steven_li
2026-06-08 13:28:41 +08:00
parent 6dc580ab26
commit a925f0e77f
3 changed files with 136 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

47
app-instance/backend/beaver/skills/learning/synthesizer.py
View File

@ -17,8 +17,9 @@ class SkillDraftSynthesizer:
evidence_packet: EvidencePacket,
provider: LLMProvider,
model: str,
base_skill: dict[str, Any] | None = None,
) -> dict[str, Any]:
return await self._synthesize(candidate, evidence_packet, provider, model, "revise")
return await self._synthesize(candidate, evidence_packet, provider, model, "revise", base_skill=base_skill)
async def synthesize_new_skill(
self,
@ -27,7 +28,7 @@ class SkillDraftSynthesizer:
provider: LLMProvider,
model: str,
) -> dict[str, Any]:
return await self._synthesize(candidate, evidence_packet, provider, model, "new")
return await self._synthesize(candidate, evidence_packet, provider, model, "new", base_skill=None)
async def synthesize_merge(
self,
@ -35,8 +36,9 @@ class SkillDraftSynthesizer:
evidence_packet: EvidencePacket,
provider: LLMProvider,
model: str,
base_skill: dict[str, Any] | None = None,
) -> dict[str, Any]:
return await self._synthesize(candidate, evidence_packet, provider, model, "merge")
return await self._synthesize(candidate, evidence_packet, provider, model, "merge", base_skill=base_skill)
async def _synthesize(
self,
@ -45,15 +47,18 @@ class SkillDraftSynthesizer:
provider: LLMProvider,
model: str,
action: str,
*,
base_skill: dict[str, Any] | None,
) -> dict[str, Any]:
prompt = self._build_prompt(candidate, evidence_packet, action)
prompt = self._build_prompt(candidate, evidence_packet, action, base_skill=base_skill)
response = await provider.chat(
messages=[
{
"role": "system",
"content": (
"You synthesize Beaver skill drafts from execution evidence. "
"Return only JSON with keys: frontmatter, content, change_reason."
"Return only JSON with keys: frontmatter, content, change_reason, "
"preserved_sections, changed_sections, dropped_sections."
),
},
{"role": "user", "content": prompt},
@ -69,11 +74,30 @@ class SkillDraftSynthesizer:
return self._fallback_payload(candidate, evidence_packet, action)
@staticmethod
def _build_prompt(candidate: SkillLearningCandidate, evidence_packet: EvidencePacket, action: str) -> str:
def _build_prompt(
candidate: SkillLearningCandidate,
evidence_packet: EvidencePacket,
action: str,
base_skill: dict[str, Any] | None = None,
) -> str:
tool_names = _coerce_string_list(evidence_packet.metadata.get("tool_names"))
tool_section = ", ".join(tool_names) if tool_names else "none observed"
selected_tool_names = _coerce_string_list(evidence_packet.metadata.get("selected_tool_names"))
selected_tool_section = ", ".join(selected_tool_names) if selected_tool_names else "none recorded"
base_section = ""
if base_skill:
base_section = (
"\n\nBase skill snapshot:\n"
f"- skill_name: {base_skill.get('skill_name')}\n"
f"- version: {base_skill.get('version')}\n"
f"- frontmatter: {json.dumps(base_skill.get('frontmatter') or {}, ensure_ascii=False, sort_keys=True)}\n"
f"- tool_hints: {base_skill.get('tool_hints') or []}\n"
f"- summary: {base_skill.get('summary') or ''}\n"
"Base skill content:\n"
f"{base_skill.get('content') or ''}\n"
"Preserve existing instructions unless the evidence requires a change. "
"If any section is changed or dropped, explain it in changed_sections or dropped_sections."
)
return (
f"Action: {action}\n"
f"Candidate kind: {candidate.kind}\n"
@ -83,11 +107,13 @@ class SkillDraftSynthesizer:
f"Run-selected tool names: {selected_tool_section}\n"
f"Task summaries:\n- " + "\n- ".join(evidence_packet.task_summaries)
+ "\n\nSession excerpts:\n" + "\n\n".join(evidence_packet.session_excerpts)
+ base_section
+ "\n\nReturn JSON only. The frontmatter object must include:"
+ "\n- description: a concise skill description"
+ "\n- tools: an explicit JSON array of exact tool names this skill needs. "
+ "Prefer called tool names when the workflow depends on them; use run-selected tool names only when clearly required. "
+ "Use [] only when no tool is required."
+ "\nThe JSON may include preserved_sections, changed_sections, and dropped_sections arrays."
)
@staticmethod
@ -111,6 +137,9 @@ class SkillDraftSynthesizer:
"frontmatter": frontmatter,
"content": content_value.strip(),
"change_reason": str(payload.get("change_reason") or ""),
"preserved_sections": _coerce_string_list(payload.get("preserved_sections")),
"changed_sections": _coerce_string_list(payload.get("changed_sections")),
"dropped_sections": _coerce_string_list(payload.get("dropped_sections")),
}
@staticmethod
@ -124,6 +153,9 @@ class SkillDraftSynthesizer:
"frontmatter": frontmatter,
"content": str(payload.get("content") or "").strip(),
"change_reason": str(payload.get("change_reason") or ""),
"preserved_sections": _coerce_string_list(payload.get("preserved_sections")),
"changed_sections": _coerce_string_list(payload.get("changed_sections")),
"dropped_sections": _coerce_string_list(payload.get("dropped_sections")),
}
@staticmethod
@ -138,6 +170,9 @@ class SkillDraftSynthesizer:
},
"content": f"# {title}\n\n## Evidence\n\n{content}\n",
"change_reason": candidate.reason or f"Fallback {action} synthesis.",
"preserved_sections": [],
"changed_sections": [],
"dropped_sections": [],
}