The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several Node.js scripts in the scripts/ directory to automate SEO tasks. For instance, render-check.mjs uses execFileSync to call the agent-browser CLI for visual audits. These executions are restricted to the skill's operational scope, such as verifying that pages render correctly on mobile and desktop viewports without broken assets.
[DATA_EXFILTRATION]: While the skill performs network operations to external APIs like Google Search Console, Serper, DataForSEO, and Firecrawl, these are documented and legitimate integrations for SEO data analysis. The skill implements strong security boundaries, instructing users to manage credentials via environment variables or secure files with 0600 permissions and emphasizing the prevention of secret exposure in workspace files.
[EXTERNAL_DOWNLOADS]: The documentation references several well-known technology services and tools, including firecrawl-cli and the author's own @jorgemenadev/superaseo package. These are standard dependencies for the skill's described purpose of SEO automation and reporting.
[INDIRECT_PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection as it processes data from search engine results and competitor web pages. However, this is mitigated by extensive instructions for human oversight and factual verification of all ingested data.
Ingestion points: External content enters the workspace via search API responses and Firecrawl scraping tasks.
Boundary markers: Processed data is isolated into structured reports within the .seo/reports/data/ directory.
Capability inventory: The skill can perform network requests to legitimate SEO providers, write to the local workspace, and execute rendering checks using an external browser tool.
Sanitization: The scripts handle data as structured JSON or plain text for manual review; the instructions require humans to look at screenshots and verify claims before publication.

seo-growth-workspace