human-review

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions include a command to install an external dependency from a third-party source not belonging to the verified vendor list. Specifically, it provides the user with npx skills@latest add mattpocock/skills --skill codebase-design to extend the skill's capabilities.
  • [INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting and analyzing untrusted external data from pull requests, which could contain adversarial instructions aimed at bypassing review logic or influencing the agent's summary.
  • Ingestion points: The skill reads PR descriptions, base-to-head diffs, linked specifications, tickets, and previous AI review results, as detailed in the 'Prepare the PR review' section of SKILL.md.
  • Boundary markers: There are no explicit instructions to use delimiters or ignore instructions embedded within the PR data.
  • Capability inventory: The skill has the capability to write local files to maintain a review record (human-review/pr-<number>.md).
  • Sanitization: The skill lacks explicit sanitization or validation steps for the content ingested from external PR artifacts before processing them or writing them to the local record.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 12:39 PM