differential-review

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's fundamental purpose is to analyze external, untrusted code changes, which exposes the agent to prompt injection attacks embedded in code comments, documentation, or string literals.\n
  • Ingestion points: Processes pull requests, commit ranges, and git diffs as core inputs for analysis.\n
  • Boundary markers: The skill lacks explicit instructions or markers to distinguish between the code being analyzed and instructions to the agent, creating a risk that the agent may interpret data as commands.\n
  • Capability inventory: The methodology involves executing git commands (git blame) and interacting with downstream reporting skills (issue-writer) that can write files.\n
  • Sanitization: The workflow does not specify any validation, filtering, or escaping of the external code content before it is processed or used to generate audit reports.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 07:57 AM
Security Audit — agent-trust-hub — differential-review