resolving-merge-conflicts
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to process data from untrusted external sources which could be used to influence the agent's behavior.
- Ingestion points: The agent is directed in
SKILL.md(Step 2) to read commit messages, check PRs, and check original issues/tickets to understand intent. - Boundary markers: Absent. There are no instructions to the agent to treat this external content as data only or to ignore embedded instructions within these sources.
- Capability inventory: The agent has the capability to execute git commands (stage, commit, rebase) and run discovered project-specific automated checks, tests, and formatters (
SKILL.mdStep 4 and 5). - Sanitization: Absent. There is no requirement for the agent to sanitize or validate the content read from commits or PRs before using it to decide on conflict resolution or running commands.
- [COMMAND_EXECUTION]: The skill requires the agent to execute shell commands to resolve conflicts and verify project integrity.
- Evidence:
SKILL.mdStep 4 ("run them, typically typecheck, then tests, then format") and Step 5 ("Stage everything and commit. If rebasing, continue the rebase process") involve shell-level interactions. While necessary for the skill's purpose, this is the surface that could be exploited by an indirect prompt injection.
Audit Metadata