The Agent Skills Directory

[DATA_EXFILTRATION]: The skill performs network operations to non-whitelisted domains to perform visual research and site audits.
The skill uses a browser tool to research competitor websites for design inspiration.
The audit command accepts a --url parameter to fetch and analyze content from arbitrary web addresses.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from both the local codebase and external websites, which could potentially contain malicious instructions.
Ingestion points: Scanning local CSS, Tailwind, and styled-components files; retrieving content from external competitor websites; and auditing user-provided URLs.
Boundary markers: There are no instructions or delimiters defined to isolate the external content or instruct the agent to ignore instructions embedded within the processed data.
Capability inventory: The skill has the ability to write multiple files to the filesystem (DESIGN.md, design-tokens.json, design-preview.html) and use browser tools for web navigation.
Sanitization: The instructions do not specify any validation, filtering, or escaping of the content retrieved from external sources before it is processed by the agent.

design-system