pr-labeling
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection Surface: The skill processes content from pull requests that may be authored by external, untrusted contributors. This content is used to determine classification labels.
- Ingestion points: The
read_pull_requesttool is used to fetch the title, description, and file patches from external PRs (SKILL.md, step 1). - Boundary markers: The instructions do not specify the use of delimiters or specific tags to help the agent distinguish between its operational instructions and the potentially adversarial content within the PR data.
- Capability inventory: The skill possesses the capability to modify repository metadata through the
add_labeltool (SKILL.md, step 5). - Sanitization: There are no explicit steps for sanitizing or filtering the PR content to remove potential injection attempts before analysis.
Audit Metadata