skills/mthines/agent-skills/create-pr/Gen Agent Trust Hub

create-pr

Pass

Audited by Gen Agent Trust Hub on May 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill retrieves untrusted content from the repository's git history (git log), code changes (git diff), and GitHub Actions logs (gh run view --log-failed) to generate pull request descriptions and triage build failures. This data can be influenced by any contributor to the repository, creating a surface for indirect prompt injection where malicious text in a commit message or a failing test's output could attempt to override agent instructions.
  • Ingestion points: Repository information is gathered via git log, git diff, and gh run view within the SKILL.md workflows.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when interpolating this retrieved data into agent prompts.
  • Capability inventory: The skill has the ability to perform git push, gh pr create, and gh run rerun. Additionally, it can delegate complex fixing tasks to a /ci-auto-fix subagent which has capabilities to modify files and commit changes.
  • Sanitization: The skill does not perform sanitization, filtering, or validation of the content retrieved from git history or CI logs before it is processed by the language model.
Audit Metadata
Risk Level
SAFE
Analyzed
May 17, 2026, 04:53 AM
Security Audit — agent-trust-hub — create-pr