precommit

Pass

Audited by Gen Agent Trust Hub on Apr 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using tools like node, npm, ruff, and cargo to automate project quality gates. It specifically attempts to run a local script located at .claude/scripts/precommit-runner.js if detected within the project.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it reads and processes untrusted project manifest files to determine its execution path.
  • Ingestion points: Project configuration files such as package.json, pyproject.toml, and the local precommit-runner.js script.
  • Boundary markers: Absent; there are no delimiters or instructions to prevent the agent from obeying instructions embedded in these project files.
  • Capability inventory: Extensive bash tool access provided in the allowed-tools frontmatter, including multiple runtimes and build systems.
  • Sanitization: The skill lacks mechanisms to validate or sanitize script names and configurations extracted from the local filesystem before they are interpolated into shell commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 20, 2026, 07:24 PM
Security Audit — agent-trust-hub — precommit