comprehensive-review-full-review

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external code provided via the $ARGUMENTS variable. This code is then interpolated into prompts for various specialized subagents (e.g., code-reviewer, security-auditor, architect-review).
  • Ingestion points: Untrusted code enters the workflow through $ARGUMENTS in the instructions within SKILL.md.
  • Boundary markers: The prompts do not utilize explicit boundary markers (like XML tags or clear delimiters) or specific 'ignore embedded instructions' warnings for the subagents.
  • Capability inventory: The skill utilizes a Task tool to invoke agents with significant analysis capabilities, which likely include file system read access to perform the reviews.
  • Sanitization: There is no evidence of input sanitization or validation of the content within $ARGUMENTS before it is passed to the sub-prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 04:07 PM
Security Audit — agent-trust-hub — comprehensive-review-full-review