The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx -y modern-web-guidance@latest to download and execute a web guidance utility from the npm registry during the setup and SEO auditing phases.
[COMMAND_EXECUTION]: The skill executes local scripts including skills/deploy/scripts/ai-readability.mjs for content visibility analysis and node skills/rollout/scripts/findings.mjs record for logging findings into a project ledger.
[DATA_EXFILTRATION]: When invoked with the --deploy flag, the skill transmits a comprehensive audit report—which includes site screenshots, brand extraction data, and technical findings—to a remote deployment server via the DA transport protocol.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted data from user-provided URLs.
Ingestion points: External website content is crawled and saved to pages/*.json and _brand-extraction.json via the stardust:extract utility.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the prompts processing the extracted site data.
Capability inventory: The skill possesses capabilities for network operations (curl, Playwright), file system writes (audit.json), and remote report deployment.
Sanitization: There is no evidence of sanitization or filtering applied to the external web content before it is analyzed by the AI or used to generate the final report.

audit