The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from Shopify article bodies and external research URLs, which is then processed by the agent to generate updates for the store. While the skill includes sanitization logic, the potential for adversarial content to influence the agent's actions exists.
Ingestion points: shopify-blog-seo-admin.mjs retrieves article data via GraphQL and external content via fetchPublic in scripts/lib/public-fetch.mjs.
Boundary markers: No explicit markers are used in the instructions to the agent to help it distinguish between instructions and ingested data.
Capability inventory: The skill has access to the Shopify Admin API for article updates, file system access for writing reports and artifacts, and shell execution capabilities via the Shopify CLI.
Sanitization: The skill employs a safeArticleHtml function to filter unsafe tags (scripts, iframes) and event handlers before performing updates or rendering reports. Generated reports also use a restrictive Content Security Policy (CSP).
[DATA_EXPOSURE_&_EXFILTRATION]: The skill performs network requests to the Shopify Admin API and various external research URLs. It also reads Shopify credentials from a local environment file.
Mitigation: The scripts/lib/public-fetch.mjs utility implements robust SSRF protection, blocking requests to private IP ranges, loopback addresses, and invalid DNS resolutions.
[DYNAMIC_EXECUTION]: The skill utilizes execFile to invoke the Shopify CLI and Node.js for environment setup and store interactions.
Mitigation: Input variables like the Shopify store domain are strictly validated against regular expressions in scripts/lib/shopify-dev-dashboard-auth.mjs to prevent command injection.

shopify-blog-seo-optimizer