make-repo-contribution
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves inspecting and modifying existing codebases and repository instructions, which creates a vulnerability to indirect prompt injection if the source materials contain adversarial instructions designed to influence the agent's behavior. \n
- Ingestion points: The workflow in
SKILL.mdrequires the agent to "Inspect the available files, runtime, versions, inputs, and existing conventions" and "repository instructions". \n - Boundary markers: The skill includes explicit boundaries, instructing the agent to "Preserve the user's... authorization boundaries" and "Keep the core Make Repo Contribution work local". It also requires user confirmation before paid API calls or repository modifications like staging and committing. \n
- Capability inventory: The agent is authorized to implement code changes, run tests, and execute external capabilities through tools such as
sandbase_run. \n - Sanitization: No specific technical sanitization methods are described for the content of files or external documentation retrieved during the discovery process.
Audit Metadata