The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses node to execute a local script orchestrate.mjs located in the ~/.claude/skills/refactor-chain/ directory. This script is part of the vendor's toolset and is used to check the status of the refactoring process.
[SAFE]: The skill manages its own state by writing a plan to .refactor-chain/plan.md and updating a local state.json file. These operations are restricted to the project's refactor-chain subdirectory and follow established tool protocols.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes external files such as spec.md and tasks.md to coordinate refactoring steps.
Ingestion points: state.json (SKILL.md), .specify/spec.md, .specify/plan.md, tasks.md (references/spec-kit-interop.md).
Boundary markers: Not explicitly defined in the prompts for these files.
Capability inventory: Subprocess execution via node orchestrate.mjs (SKILL.md).
Sanitization: No explicit sanitization of file content before processing is mentioned.

refactor-plan-gate