skills/dmmulroy/skills/code-review/Gen Agent Trust Hub

code-review

Pass

Audited by Gen Agent Trust Hub on Jul 7, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill is strictly read-only and explicitly instructs the agent not to modify files or apply patches during the review process.
  • [SAFE]: File access is limited to the repository context and a specific standards folder; no access to sensitive system directories is requested or attempted.
  • [COMMAND_EXECUTION]: The skill utilizes git commands to identify review targets and diffs. This is standard functionality for repository-integrated agents and does not involve arbitrary user-input execution.
  • [PROMPT_INJECTION]: The skill processes untrusted code (the review target), creating a surface for indirect prompt injection.
  • Ingestion points: The skill reads code from user-specified files, git diffs, and branch comparisons (SKILL.md, Section 1).
  • Boundary markers: The instructions do not define specific delimiters or warnings to isolate the reviewed code from the agent's instructions.
  • Capability inventory: The skill is restricted to read-only operations and has no access to external networks, which prevents data exfiltration or system modification via injection.
  • Sanitization: No content filtering or sanitization is performed on the code being reviewed.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 7, 2026, 03:16 PM
Security Audit — agent-trust-hub — code-review