The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses npx to run the impeccable CLI for design scanning and to manage a local live detection server.
[EXTERNAL_DOWNLOADS]: The skill fetches the impeccable package from the official npm registry during execution if it is not already present.
[REMOTE_CODE_EXECUTION]: To provide visual feedback, the skill injects a script (detect.js) into the browser environment from a locally-hosted server started by the agent.
[PROMPT_INJECTION]: The skill processes untrusted content from source files and browser console logs, representing an indirect prompt injection surface.
Ingestion points: Reads local source code files (HTML, JSX, TSX, etc.) and monitors browser console logs via the read_console_messages tool.
Boundary markers: The skill instructs the agent to filter console output using the [impeccable] prefix to distinguish tool results from other logs.
Capability inventory: The skill possesses the ability to execute shell commands, inject browser scripts, and read local project files.
Sanitization: Content is passed through a synthesis step where the agent is instructed to weave the findings into a report, providing a layer of cognitive filtering before final output.

critique