resolving-merge-conflicts
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of natural language instructions for git conflict resolution. It does not include any scripts, commands that perform network exfiltration, or unauthorized file access. The instructions follow standard software development practices such as running automated tests and linters before committing changes.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze external, potentially untrusted data to resolve conflicts.
- Ingestion points: Git commit messages, Pull Request descriptions, and issue/ticket comments (SKILL.md).
- Boundary markers: None present in the instructions.
- Capability inventory: Git command execution and project-specific testing/formatting tools (SKILL.md).
- Sanitization: None specified; the agent relies on its own ability to interpret the intent of the conflict sources.
Audit Metadata