contributions-implement
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external repositories and GitHub pull request reviews, creating a surface for indirect prompt injection attacks.
- Ingestion points: Reads repository documentation (CONTRIBUTING.md, AGENTS.md, CLAUDE.md), issue/PR templates, and maintainer review comments via GitHub APIs.
- Boundary markers: The instructions do not specify the use of delimiters or "ignore embedded instructions" warnings when interpolating this untrusted data into the agent's context.
- Capability inventory: The skill has the ability to execute Bash commands via git, gh, and multica, allowing for file system modifications and network interactions.
- Sanitization: No automated sanitization or verification of external content is described before it is used to influence agent actions or generate commands.
- [COMMAND_EXECUTION]: The skill is designed to run arbitrary build, test, and linting commands (e.g., make lint, go test) found within the target repository's configuration files. This results in the execution of code defined by potentially untrusted projects.
Audit Metadata