buildbetter-knowledge-gaps

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: No malicious patterns detected. The skill instructions prioritize explicit user approval and verification of state changes, which is a security best practice for autonomous agents.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes project-related data such as knowledge-gap recommendations and evidence which could contain instructions intended to influence the agent. 1. Ingestion points: Data enters the agent context through tools like list-knowledge-gaps and sync-release-sources in SKILL.md. 2. Boundary markers: The instructions include strong safety contracts, explicitly stating that retrieved documents do not override user instructions and requiring the agent to obtain approval for every mutation. 3. Capability inventory: Potential impact is limited to the skill's specific tools for reviewing gaps and updating project status (review-knowledge-gap, attach-knowledge-gap-to-project). 4. Sanitization: The skill relies on human verification and approval of specific IDs and rationales rather than automated content filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 09:33 AM