stage-review

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates git operations and the execution of project-specific verification commands (e.g., tests, linters) defined in local repository configuration files like package.json, Makefile, or justfile to ensure code quality before staging.
  • [COMMAND_EXECUTION]: The bundled script scripts/stage_review.py invokes subprocess.run to manage git staging and committing. It follows security best practices by passing arguments as a list and validating that file paths remain within the repository boundaries.
  • [PROMPT_INJECTION]: The workflow relies on external configuration files to determine which commands to run, creating an indirect prompt injection surface where a malicious repository could specify dangerous commands for the agent to execute.
  • Ingestion points: Files like .no-mistakes.yaml, package.json, Makefile, and justfile (SKILL.md)
  • Boundary markers: None identified
  • Capability inventory: Execution of shell commands via the agent and git automation via scripts/stage_review.py (scripts/stage_review.py)
  • Sanitization: Path validation is performed in the Python script, but no sanitization is specified for command definitions in repository config files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 10:24 PM
Security Audit — agent-trust-hub — stage-review