pr-creator

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs shell command execution using standard repository management tools. Evidence: It executes git for repository state and diffing, as well as gh and glab for interacting with GitHub and GitLab hosting platforms.
  • [DATA_EXFILTRATION]: This skill transmits repository data, such as branch diffs and commit history, to external hosting providers (GitHub, GitLab, Bitbucket). This behavior is documented, aligns with the primary purpose of creating pull requests, and targets well-known services.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it uses repository content (diffs and logs) to draft PR descriptions. Evidence: (1) Ingestion points: Data is read via git log and git diff in SKILL.md. (2) Boundary markers: No delimiters or ignore instructions are applied to the ingested diff content. (3) Capability inventory: The skill can push branches and create pull requests. (4) Sanitization: The risk is mitigated by a mandatory human-in-the-loop preview and confirmation step (Step 7), ensuring user oversight of all generated content.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 04:20 PM