lovstudio-gh-contribute
Pass
Audited by Gen Agent Trust Hub on May 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands via
gitand theghCLI to perform repository management tasks. - Evidence: Multiple steps in
SKILL.mdinvolve commands likegit status,git diff,git push, andgh pr createto manage the lifecycle of a pull request. - [EXTERNAL_DOWNLOADS]: The skill fetches documentation and metadata from GitHub repositories to ensure contributions follow project guidelines.
- Evidence: Step 2 in
SKILL.mduses the GitHub API to download content such asCONTRIBUTING.mdand PR templates from the target upstream repository. - [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it processes instructions found in external repository files.
- Ingestion points: Fetching
CONTRIBUTING.md,CODE_OF_CONDUCT.md, and PR templates from untrusted upstream repositories in Step 2 ofSKILL.md. - Boundary markers: No specific delimiters are used when interpreting the content of these external files.
- Capability inventory: The agent can perform
git commit,git push, andgh pr createbased on its interpretation of the repo rules. - Sanitization: There is no explicit sanitization of the fetched repository documentation before the agent processes it.
Audit Metadata