utility-pm-skill-auditor

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local shell script (bash scripts/pre-tag-validate.sh) to run validation tools as part of the audit process. This is the primary mechanism for enforcing repository standards.\n- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface where it reads the contents of an external file (subagents/pm-skill-auditor.md) and treats the text as system instructions for the current session. This behavior is the intended functional mechanism for dispatching logic on non-Claude clients.\n
  • Ingestion points: subagents/pm-skill-auditor.md (operating instructions) and docs/internal/release-plans/v2.16.0/spec_pm-skill-auditor.md (check catalog).\n
  • Boundary markers: Absent. The agent is instructed to execute the file body directly.\n
  • Capability inventory: The agent can read files across the repository and execute bash scripts via scripts/pre-tag-validate.sh.\n
  • Sanitization: Absent. The skill relies on the integrity of the local repository files.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 12:33 AM
Security Audit — agent-trust-hub — utility-pm-skill-auditor