The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes standard Ruby development utilities such as rubocop, brakeman, rspec, erblint, and yard using bundle exec for code analysis, testing, and documentation generation.
[EXTERNAL_DOWNLOADS]: Fetches vulnerability database updates using bundle-audit check --update, which is a standard procedure for identifying insecure dependencies from established Ruby security sources.
[INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection because it ingests and interprets untrusted source code to perform analysis and refactoring.
Ingestion points: Reads files from app/controllers, app/models, app/services, app/jobs, and spec/ as defined in the Phase 1 instructions.
Boundary markers: No specific delimiters or instructions are used to distinguish between source code and potential malicious instructions embedded in code comments.
Capability inventory: Shell execution of multiple CLI tools and the ability to modify project files during the refactoring phase.
Sanitization: No input validation or sanitization of the processed source code is performed before analysis.

quality