vp-github
Pass
Audited by Gen Agent Trust Hub on Nov 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides templates for processing local files (images, videos, research documents, and HTML) for upload to GitHub. * Ingestion points: Local files and user-supplied metadata are processed across references/attachments.md and references/issue-artifacts.md. * Boundary markers: Code templates do not include specific delimiters, but the instructions explicitly warn the user to treat downloaded documents as untrusted input. * Capability inventory: Capabilities include the GitHub CLI (gh), curl for network operations, and JavaScript snippets for browser automation. * Sanitization: No automated sanitization is provided in the templates; the skill relies on manual verification steps described in the documentation.
- [COMMAND_EXECUTION]: The skill provides various bash command templates using the GitHub CLI (gh) and curl to interact with GitHub APIs and upload media. These commands are consistent with the skill's stated purpose of managing repository artifacts and attachments.
- [SAFE]: The skill demonstrates safe credential handling by advising the use of curl --config
- to prevent tokens from appearing in process argument lists. All external references and download links point to official GitHub domains or the author's own repository, representing standard development workflows.
Audit Metadata