resolving-merge-conflicts
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest external data from PRs, issues, and commit messages. This creates a surface where an attacker could potentially embed instructions in project metadata to influence the agent's behavior during conflict resolution.
- Ingestion points: Git history, commit messages, PRs, and issues (SKILL.md).
- Boundary markers: None present.
- Capability inventory: Command execution (git, compiler, test suite, linter) and file modification via staging/committing.
- Sanitization: None observed.
- [COMMAND_EXECUTION]: The instructions require running the local project's compiler, test suite, and linter to verify the conflict resolution. This involves executing commands within the project environment, which is consistent with the skill's intended purpose.
Audit Metadata