skills/idaibin/skills/repo-review/Gen Agent Trust Hub

repo-review

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes shell commands for repository inspection, specifically git status --short, git diff --stat, and git diff --name-status. It also allows for 'clean generation' commands in the protocol-contract profile. These are restricted to non-mutating checks, and the skill explicitly forbids editing files, staging changes, or modifying Git configuration.
  • [PROMPT_INJECTION]: The skill ingests untrusted data from the repository being reviewed, such as guidance files (AGENTS.md, CLAUDE.md) and commit metadata, which could serve as a surface for indirect prompt injection. This is mitigated by the skill's structured workflows and the requirement to verify all finding facts against an immutable basis.
  • Ingestion points: Repository guidance files, source code, diff stat outputs, and pull request metadata.
  • Boundary markers: The skill defines explicit review bases (Worktree vs. Fixed-basis) and independent review axes (Standards and Spec) to maintain evidence integrity.
  • Capability inventory: File system read access and Git CLI status/diff operations.
  • Sanitization: Instructions mandate independent verification of facts and specifically warn against inferring intent or specifications from the code diff itself.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 03:40 AM
Security Audit — agent-trust-hub — repo-review