The Agent Skills Directory

[COMMAND_EXECUTION]: Automates project analysis using pnpm, tsc, and grep for linting and pattern matching.
[COMMAND_EXECUTION]: Employs node -e in Step 3 to execute a JavaScript script for line-count calculations.
[COMMAND_EXECUTION]: Uses bash and PowerShell scripting in Step 8 to find dead code and unused files.
[PROMPT_INJECTION]: Identifies an indirect prompt injection surface (Category 8) where project file data is processed by the agent.
Ingestion points: Project file names and source code content retrieved from the file system.
Boundary markers: None; file content is not isolated using specific markers or delimiters.
Capability inventory: The skill utilizes Shell, Read, Grep, and Glob tools to inspect the environment.
Sanitization: No explicit sanitization of file-based data is performed prior to inclusion in shell command executions.

code-quality