create-pr
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes repository content, including commit messages, branch names, and configuration files, to automatically generate pull request titles and descriptions. This creates a surface for indirect prompt injection if an attacker can influence the repository's history or local files. However, the impact is low as the output is restricted to the PR metadata and the skill explicitly forbids executing external commands or installations during these checks.
- Ingestion points: Git commit history (
git log), diffs (git diff), and local configuration files (e.g.,drizzle.config.ts,prisma/schema.prisma). - Boundary markers: Not explicitly defined in the instructions.
- Capability inventory: Git operations (
git push), GitHub CLI operations (gh pr create,gh pr edit), and GitLab CLI operations (glab mr create). - Sanitization: None specified, but the generated output is reviewed/edited by the user through the PR process.
Audit Metadata