create-pr
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a theoretical attack surface because it reads external repository data to draft text content.
- Ingestion points: The instructions require the agent to read
git diff,git log, and external ticket descriptions withinSKILL.md. - Boundary markers: Absent. There are no explicit instructions for the agent to ignore potential prompt injections embedded within commit messages or ticket content.
- Capability inventory: The skill uses the
ghandglabCLI tools to create and modify pull requests as described inSKILL.md. - Sanitization: The process mandates a "secrets check" and a "humanizer pass" in Step 4 to ensure the final output is safe and accurate.
- [COMMAND_EXECUTION]: The skill executes command-line tools to manage the pull request lifecycle.
- Evidence: The agent uses
gh pr create,gh pr edit, andglab mr createto push changes and open draft PRs. These operations are consistent with the skill's primary purpose and utilize appropriate safety flags like--draft.
Audit Metadata