The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local Node.js and Bash scripts to perform SEO audits. These scripts are invoked with user-supplied URLs and paths to analyze site structure and metadata.
[DATA_EXFILTRATION]: The skill performs network requests to external domains provided by the user to retrieve public SEO data. These operations are restricted to HTTP GET/HEAD requests for auditing and do not transmit sensitive local data.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted data from external websites (e.g., meta tags, JSON-LD) which then informs automated code modifications in the user's project.
Ingestion points: Content is retrieved from remote URLs via fetch in scripts/audit-meta.mjs and scripts/extract-jsonld.mjs, and via curl in scripts/redirect-trace.sh.
Boundary markers: Output from analysis scripts is provided to the agent without specific delimiters or instructions to ignore embedded commands.
Capability inventory: The skill includes instructions for the agent to modify the project's codebase (Phase 3 — fix) based on the ingested analysis results.
Sanitization: While the scripts parse HTML and JSON into structured segments, the resulting text fields are not sanitized against malicious instructional content. This represents a low-risk surface inherent to the tool's primary purpose.

codebase-seo