The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/test_skill.py utility uses the subprocess module to execute the local scripts/audit_lean_text.ts script via the bun runtime. This execution is used solely for functional validation and does not invoke a shell, which minimizes the risk of command injection.
[SAFE]: The skill provides instructions and tools to minimize filler text in web development based on guidance from well-known sources such as the W3C and Nielsen Norman Group. It does not perform any network operations or access sensitive system credentials.
[SAFE]: The skill performs static analysis on HTML and JSX files to ensure accessibility standards are met. While this processing of external source code represents a theoretical surface for indirect prompt injection, it is managed within a limited diagnostic context.
Ingestion points: Source files are read locally by scripts/audit_lean_text.ts for regex-based analysis.
Boundary markers: Content is processed directly as text; no special delimiters are required for the audit logic.
Capability inventory: The skill enables the agent to audit markup and suggest refactors; it does not have autonomous access to high-privilege system operations.
Sanitization: The audit script only outputs diagnostic reports and does not execute or evaluate any of the code it analyzes.

lean-text-scaffolding