domains / families / business / enterprise-artifact-search
enterprise-artifact-search
separate
Multi-hop evidence search + structured extraction over enterprise artifact datasets (docs/chats/meetings/PRs/URLs). Strong disambiguation to prevent cross-product leakage; returns JSON-ready entities plus evidence pointers.
When to use
- multi-hop evidence gathering
- retrieving IDs, names, dates, and roles from artifacts
- searching across docs, slack, meetings, PRs, and URLs
- extracting structured entities like employee IDs and doc IDs
- verifying product scope to reject distractors
- keeping context lean by delegating to a subagent
When not to use
- inferring answers without artifact retrieval
- using oracle or label fields like ground_truth
- loading large files directly into main agent context
Skill page
Source
| Repository | Path |
|---|
| benchflow-ai/skillsbench | tasks/enterprise-information-search/environment/skills/enterprise-artifact-search/SKILL.md |