The Agent Skills Directory

[SAFE]: The skill provides structured guidance for a specific developer workflow (Rails engine extraction). No malicious patterns, data exfiltration, or obfuscation techniques were identified across any of the files.
[COMMAND_EXECUTION]: The skill includes instructions for common file operations (mv, mkdir) and testing (bundle exec rspec). These are standard development activities and are used appropriately within the context of the skill's purpose to guide an agent through refactoring.
[INDIRECT_PROMPT_INJECTION]: As a code transformation tool, the skill processes user-provided tasks and existing source code. While this constitutes an attack surface, the instructions are scoped to a specific refactoring procedure with no indication of unsafe handling of external data.
Ingestion points: User tasks (prompts) and existing Rails source code files.
Boundary markers: Uses Markdown sections and code blocks to separate instructions from examples.
Capability inventory: File manipulation (mv, mkdir) and test execution (bundle exec rspec) within a Rails project context.
Sanitization: Standard process adherence is enforced by the "HARD-GATE" instruction to preserve behavior before refactoring.

extract-engine