merge-resolve
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local diagnostic bash script
scripts/gather_conflicts.shto analyze the state of git conflicts, including unmerged files and commit history. - [COMMAND_EXECUTION]: The agent is instructed to run project-specific verification tools such as linters or type-checkers (e.g.,
npm run lint,tsc --noEmit) to ensure conflict resolution does not introduce regression. - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes untrusted data from the repository.
- Ingestion points: File contents, git conflict markers, and commit messages (retrieved via
git diffandgit loginscripts/gather_conflicts.sh). - Boundary markers: The skill uses a structured markdown plan for resolutions but does not include explicit instructions to the agent to ignore potentially malicious directions embedded in the code or commit history being processed.
- Capability inventory: The skill utilizes
git,bash, and project-defined linting/testing utilities. - Sanitization: No explicit sanitization or filtering of the repository-sourced data is performed before analysis.
Audit Metadata