code-review

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes Git commands such as git diff, git log, and git rev-parse using user-provided references. It also instructs the agent to run a local setup script, /setup-matt-pocock-skills, if necessary documentation is missing from the environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from the repository and interpolates it into prompts for specialized sub-agents.
  • Ingestion points: Git diff outputs, commit messages, and content from specification files found in the docs/, specs/, or .scratch/ directories are ingested into the sub-agent context as described in the Process section of SKILL.md.
  • Boundary markers: Absent. The instructions for spawning the sub-agents do not specify the use of delimiters or warnings to ignore instructions embedded within the diff or specification content.
  • Capability inventory: The agent has the capability to read files and execute Git commands within the repository environment.
  • Sanitization: No sanitization or filtering of the Git output or specification content is performed before passing the data to the sub-agents.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 05:44 PM
Security Audit — agent-trust-hub — code-review