ai-agent-pr-metadata

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes data from GitHub PR comments and commits, creating a surface for indirect prompt injection.
  • Ingestion points: PR review comments, issue comments, and commit history are fetched using gh api from $GITHUB_REPOSITORY and stored in temporary files.
  • Boundary markers: The instructions do not specify explicit delimiters or "ignore" instructions for the ingested content when passed to the verification script.
  • Capability inventory: The skill can execute shell commands, run local Python scripts, and modify PR labels and comments.
  • Sanitization: Sanitization logic is not explicitly defined in the provided instruction set; validation is delegated to the verify-pr-governance.py script.
  • [COMMAND_EXECUTION]: The skill executes various command-line tools for PR management and governance verification.
  • Evidence: Use of gh api, gh pr comment, gh pr checks, and ocr rules check within shell scripts.
  • Evidence: Invocation of a local Python script support/ai-engineering-workflow/scripts/verify-pr-governance.py.
  • [SAFE]: The skill incorporates security best practices by mandating the redaction of sensitive information from public records.
  • Evidence: Rules explicitly forbid including credential values or local file paths in PRs or comments, requiring the use of [redacted] placeholders.
  • [SAFE]: Interaction with the Alibaba Code Review tool and GitHub CLI is conducted neutrally as part of expected developer workflows.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 10:08 AM
Security Audit — agent-trust-hub — ai-agent-pr-metadata