The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill audits external websites defined by the user in the ${var} variable, treating their content as data to be analyzed and reported.
Ingestion points: External website metadata (titles, meta descriptions, H1 tags, etc.) fetched from origins provided in the target sites variable.
Boundary markers: No specific delimiters or "ignore instructions" warnings are mentioned for the interpolation of external data into the notification digest or the durable FIXES.md report.
Capability inventory: The skill uses node to execute the audit logic, shell redirection (> and >>) to write snapshots and remediation documents to the filesystem, and the ./notify script to transmit findings.
Sanitization: The instructions do not specify any escaping or validation of the external content before it is included in the agent's reporting context.
[DYNAMIC_EXECUTION]: The skill uses a Node.js one-liner to merge multiple JSON audit result files into a single snapshot.
Evidence: node -e 'const fs=require("fs"); const sites=process.argv.slice(1).map(f=>JSON.parse(fs.readFileSync(f))); console.log(JSON.stringify({date:"${today}",run:process.env.STAMP,sites},null,2))' $files in SKILL.md step 5.

seo-audit