utility-pm-critic

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it reads external, potentially untrusted documents to perform adversarial reviews.
  • Ingestion points: As described in SKILL.md, the skill reads target PM artifacts directly from user-supplied paths via $ARGUMENTS or from the most recent artifact in the active session context.
  • Boundary markers: No specific delimiters or protective encapsulation strategies are implemented to isolate the content of the untrusted artifact from the agent's operating instructions.
  • Capability inventory: The skill possesses file-reading capabilities and generates markdown and YAML output; it does not request high-risk tool authorizations, shell execution, or network privileges.
  • Sanitization: There is an absence of textual sanitization, formatting constraints, or escaping mechanisms applied to the ingested content before it is processed by the model's instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 06:21 AM
Security Audit — agent-trust-hub — utility-pm-critic