resolving-merge-conflicts
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read external data such as pull requests (PRs), issues, and commit messages to understand the intent behind conflicting changes. This represents a potential surface for indirect prompt injection if those external sources contain instructions designed to manipulate the agent's behavior.
- Ingestion points: Git history, pull request descriptions, and issue trackers.
- Boundary markers: No specific delimiters or safety instructions are provided to the agent regarding the handling of this external content.
- Capability inventory: The agent is authorized to modify the local filesystem (resolving hunks) and execute local project commands (tests, typecheck, format).
- Sanitization: No explicit sanitization or filtering of the external text is mentioned.
- [COMMAND_EXECUTION]: The instructions direct the agent to discover and run the project's 'automated checks' (e.g., tests, typecheck, formatters). This is standard behavior for a development-focused skill and relies on the project's pre-existing configuration.
- [NO_CODE]: The skill consists entirely of natural language instructions and does not bundle any executable scripts or binary files, significantly reducing the direct attack surface.
Audit Metadata