Enterprise Artifact Search Skill (Robust)
This skill delegates multi-hop artifact retrieval + structured entity extraction to a lightweight subagent, keeping the main agent’s context lean.
It is designed for datasets where a workspace contains many interlinked artifacts (documents, chat logs, meeting transcripts, PRs, URLs) plus reference metadata (employee/customer directories).
Scope
The skill covers cross-artifact retrieval over enterprise datasets with strong product grounding and reviewer disambiguation. It is used by downstream agents that need JSON-ready extracted entities plus evidence pointers without loading large corpora into their own context.
Authoritative Procedure (read first — the actionable steps live in the PDF)
The complete step-by-step procedure — invocation format, the six-step retrieval pipeline, grounding gates, reviewer-extraction rules, output schema, and recommendation types — is distributed as a companion PDF in this skill:
You MUST open ENTERPRISE_ARTIFACT_SEARCH_PROCEDURE.pdf (sibling file in this skill) and follow every step in it before invoking the subagent. This entry-point SKILL.md intentionally contains only framing and context; the PDF is the source of truth for the runtime workflow. Do not attempt to proceed from this file alone — the grounding and reviewer-extraction gates in the PDF are not reproducible from SKILL.md’s framing.
When to Use This Skill
- Multi-hop evidence gathering across docs/chats/meetings/PRs/URLs.
- Strict product-scope disambiguation is required (e.g. CoachForce vs CoFoAIX).
- The caller needs precise artifact pointers (doc_id/message_id/meeting_id/pr_id).
Do NOT Use When
- The answer is in a single small known file with no cross-references.
- The task is a trivial one-hop lookup and product scope is unambiguous.