hf-release-notes
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- Indirect Prompt Injection Surface: The skill processes Pull Request metadata, such as titles and bodies, which are originally authored by external users on GitHub. This content is used to generate summaries and highlights for the final release notes. Because this data comes from an untrusted source, it could potentially contain instructions aimed at influencing the agent's behavior or the content of the generated notes. (1) Ingestion points: PR metadata loaded from PR JSON files in the temporary directory as described in SKILL.md and VALIDATE.md. (2) Boundary markers: The instructions do not explicitly define strict delimiters or include warnings for the agent to ignore instructions embedded within the PR body or title. (3) Capability inventory: The skill allows the agent to fetch documentation from the vendor site and write release note files to the local file system. (4) Sanitization: No specific sanitization or validation logic is specified for the text extracted from the PR files.
- External Resource Access: The skill fetches documentation pages from the official vendor documentation site to verify feature details and extract links. Evidence: SKILL.md references fetching content from the vendor's documentation index and subpages.
Audit Metadata