gm-hop
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill executes specific 'disciplines' and 'closure generators' identified by metadata labels and code comments. This dynamic dispatch is the core mechanism for its refactoring operations.\n- [COMMAND_EXECUTION]: The skill performs filesystem and version control operations using
git,codesearch, and custom dispatch commands to manage the refactoring workflow.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted code surfaces and rhetoric strings which could contain malicious instructions. Evidence chain:\n - Ingestion points: Code files specified by 'surface' and 'rhetoric' payload arguments.\n
- Boundary markers: None identified in prompt interpolation.\n
- Capability inventory: Includes file writes via 'fs_write' and version control commits via 'git_commit'.\n
- Sanitization: Not explicitly performed, though risks are mitigated by a mandatory verification subagent that must re-witness all edits independently.
Audit Metadata