Skills
Navigate
skills/labring/fastgpt/pr-review/Gen Agent Trust Hub

pr-review

Warn

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill instructs the agent to execute pnpm install and pnpm test on code retrieved from a Pull Request. This is a vector for arbitrary code execution if the PR contains malicious lifecycle scripts or tests.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection via Pull Request data.
  • Ingestion points: PR metadata and diffs fetched via gh pr view and gh pr diff in SKILL.md.
  • Boundary markers: None; instructions do not isolate untrusted PR content from the agent's logic.
  • Capability inventory: The agent has the authority to approve/request changes on Pull Requests and post comments via gh api and gh pr review.
  • Sanitization: None; the agent processes raw output from the repository tools.
  • [COMMAND_EXECUTION]: The skill makes extensive use of shell commands via git, gh, and pnpm to manage the review workflow, perform builds, and interact with the GitHub API.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 16, 2026, 12:27 AM
Security Audit — agent-trust-hub — pr-review