The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes numerous local JavaScript files (e.g., live.mjs, live-server.mjs, pin.mjs) to manage its design workflows.
[COMMAND_EXECUTION]: A management script (scripts/pin.mjs) creates new skill definition files within the agent's configuration directories (like .claude/skills or .cursor/skills), effectively altering the agent's available command set through persistence.
[COMMAND_EXECUTION]: Maintenance and setup scripts perform file deletions (rmSync, unlinkSync) and project file injections (live-inject.mjs) to manage older versions and enable the live preview mode.
[EXTERNAL_DOWNLOADS]: The critique and audit commands invoke npx impeccable, which triggers the download and execution of a package from the public NPM registry.
[DATA_EXFILTRATION]: The included local helper server (scripts/live-server.mjs) provides a /source endpoint capable of reading any file within the current project directory. Although protected by a unique token and restricted to local connections, it creates a potential surface for sensitive file exposure (e.g., .env) if the token is accessed by other scripts on the page.
[COMMAND_EXECUTION]: The skill uses git check-ignore via shell execution to identify files that should not be modified.

impeccable