pr-to-main
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from the project's git history to draft pull request descriptions.
- Ingestion points: The outputs of
git logandgit diffcommands in SKILL.md are used as context for drafting text. - Boundary markers: No delimiters or explicit instructions are used to prevent the agent from obeying instructions embedded within commit messages.
- Capability inventory: The skill has the capability to write data to a remote repository via the
gh pr createcommand. - Sanitization: No programmatic sanitization is present, although Step 3 explicitly requires the agent to ask the user to review the draft before submission.
Audit Metadata