workflow-review

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of repository files, diffs, and pull request content during the review process.
  • Ingestion points: The agent is instructed to inspect changed files, callers, tests, and data boundaries within a repository (SKILL.md).
  • Boundary markers: Absent. The instructions do not specify the use of delimiters or warnings to ignore instructions embedded within the reviewed code.
  • Capability inventory: The skill is strictly read-only. It explicitly prohibits fixes, commits, remote comments, and write ownership (SKILL.md).
  • Sanitization: Absent. There is no requirement for the agent to sanitize or filter the repository content before analysis.
  • [SAFE]: The skill includes explicit security constraints, stating that a review request is read-only and does not authorize modifications, credential acquisition, or state changes (SKILL.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 04:08 PM
Security Audit — agent-trust-hub — workflow-review