ceo-council

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local project files and git history to provide context for expert sub-agents, which could potentially allow instructions embedded in those files to influence sub-agent outputs.
  • Ingestion points: The skill reads project documentation (CLAUDE.md, README.md), domain context files, and recent git history in Step 1 and Step 3 of SKILL.md.
  • Boundary markers: The prompt template used for sub-agents in Step 4 does not utilize specific delimiters to separate the raw project data from the instructions provided to the sub-agent.
  • Capability inventory: The skill utilizes the Task tool to launch general-purpose sub-agents and has the ability to write synthesis reports to the local file system (docs/ folder).
  • Sanitization: No validation or sanitization of the gathered project context is performed before it is passed to the sub-agents.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 11:49 AM
Security Audit — agent-trust-hub — ceo-council