resolving-merge-conflicts
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read external, untrusted content which could contain malicious instructions intended to influence the agent's behavior.
- Ingestion points: Step 2 in
SKILL.mdinstructs the agent to read commit messages, PR descriptions, and issue/ticket contents. - Boundary markers: None provided; the agent is not explicitly told to ignore embedded instructions in these sources.
- Capability inventory: The skill uses shell command capabilities to perform git operations (steps 1, 5) and run automated project checks (step 4).
- Sanitization: No sanitization or validation of the external content is performed before processing.
- [COMMAND_EXECUTION]: The skill requires the execution of shell commands to interact with git and run arbitrary automated checks.
- Evidence: Step 4 in
SKILL.mddirects the agent to "discover" and run automated checks such as typecheck, tests, and formatting scripts. While necessary for the skill's purpose, this involves executing project-defined code.
Audit Metadata