The Agent Skills Directory

[SAFE]: The skill performs legitimate accessibility auditing using trusted tools. No evidence of malicious intent, obfuscation, or unauthorized access was detected.
[COMMAND_EXECUTION]: The skill runs pnpm exec eslint to perform static analysis on component source files. This is a common and expected task in a development workflow, restricted to the local project environment.
[DATA_EXFILTRATION]: Access is limited to project-related files (source code, CSS, and tests) required for the accessibility audit. No sensitive system files or credentials are accessed, and no data is transmitted to external servers.
[PROMPT_INJECTION]: The skill audits component source code, which represents an indirect prompt injection surface.
Ingestion points: Component source code, stylesheets, and test files in projects/core/src/<component>/.
Boundary markers: None present; the skill does not use specific delimiters or instructions to ignore embedded content in audited files.
Capability inventory: Local shell execution (pnpm exec eslint), file system read, and grep search.
Sanitization: None observed; the skill processes the files as provided for linting and pattern matching.

accessibility