create-pr
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill uses standard developer tools including the GitHub CLI (
gh) andgitfor their intended purposes in a development workflow. It includes safety constraints such as preventing direct pushes to protected branches and requiring confirmation for inferred issue links. - [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from the local repository and external issue trackers.
- Ingestion points: Reads repository pull request templates (
.github/pull_request_template.md), git commit logs (git log), and issue data via Linear MCP. - Boundary markers: Employs user confirmation steps for inferred issues to prevent unauthorized linking.
- Capability inventory: Uses
git pushandgh pr createto modify remote repository state. - Sanitization: Does not explicitly sanitize commit messages or issue titles before including them in the pull request body.
Audit Metadata