grimoire-refactor

Pass

Audited by Gen Agent Trust Hub on Nov 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes repository source code and external evidence records to perform refactoring tasks. Maliciously crafted code or documentation within the target repository could potentially influence the agent's logic or execution plan.
  • Ingestion points: Source code navigation through grimoire-map, audit requests, and external coordination contracts (e.g., coordination-contract.md).
  • Boundary markers: The skill instructs the agent to "inspect actual diffs" and "revalidate referenced owner evidence," but lacks explicit delimiters or instructions to ignore embedded prompts within the processed source code.
  • Capability inventory: The skill has permission to perform "reversible in-scope internal rewrites without per-batch approval," create branches/worktrees, and execute testing/build suites via grimoire-test and grimoire-check.
  • Sanitization: No specific sanitization or filtering of external source content is mentioned before it is processed by the coordination logic.
  • [COMMAND_EXECUTION]: The skill invokes several external utilities and scripts to perform its duties, including git (for worktree and branch management), cloc or tokei (for line counting and measurement), and project-specific build, lint, and integration test scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 11, 2026, 11:06 AM