The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute various Python scripts located in the local hidden directory ~/.claude/skills/ for validation tasks (e.g., ux_audit.py, security_scan.py, lint_runner.py). These commands are triggered automatically after completing work.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided code and external project files to apply coding standards and run verification scripts, creating a surface for indirect instructions to influence agent behavior.
Ingestion points: The agent reads project-wide directories (.) and specific source files (e.g., UserService.ts) to perform edits and audits.
Boundary markers: There are no explicit instructions or delimiters defined to help the agent distinguish between its instructions and content found within the analyzed files.
Capability inventory: The skill utilizes Read, Write, and Edit tools, and executes shell commands via python for its verification scripts.
Sanitization: No sanitization or validation logic is defined for the content being processed or the arguments passed to the verification scripts (e.g., the <url> argument in the Lighthouse audit).

clean-code