resolving-merge-conflicts

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The instructions follow standard software development practices for conflict resolution and do not include any malicious patterns or risky commands.
  • [COMMAND_EXECUTION]: The skill suggests running a project's existing automated checks (tests, typecheck, format). This is standard functionality for a coding agent and does not introduce new risks.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze external data sources for context. * Ingestion points: Commit messages, PR descriptions, and issue tickets are read to understand the intent of conflicting changes (SKILL.md). * Boundary markers: None explicitly defined; the agent is instructed to use this data for context only. * Capability inventory: The skill utilizes git commands and local project testing tools (SKILL.md). * Sanitization: None defined; the agent relies on its internal safety filters when processing these texts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 10:09 PM
Security Audit — agent-trust-hub — resolving-merge-conflicts