The Agent Skills Directory

[SAFE]: The skill operates within a structured framework for SEO optimization, focusing on technical health, keyword mapping, and content authority. It references industry-standard tools and metrics (e.g., Google Search Console, Ahrefs, Core Web Vitals) and adheres to white-hat SEO principles such as E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness).
[SAFE]: The skill includes a dedicated references/anti-patterns.md file that explicitly forbids malicious or deceptive SEO tactics, including link buying, hidden text, keyword stuffing, doorway pages, and cloaking. This demonstrates a clear alignment with platform safety guidelines.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data such as website URLs (store_url), search console metrics, and competitor data. This constitutes an indirect prompt injection surface. However, the risk is mitigated by explicit boundary markers in the workflow:
Ingestion points: Data enters via user-provided URLs and tool-generated summaries (GSC, Analytics, 3rd party SEO tools).
Boundary markers: The Phase 4: Output & Verification step and the anti-patterns.md reference file provide constraints to prevent the agent from executing malicious instructions embedded in processed web content.
Capability inventory: The skill has the capability to write to the brand-brain directory (e.g., products.md, learnings.jsonl). It does not request network tools or subprocess execution in its frontmatter.
Sanitization: The skill uses a structured completion protocol and 90-day Action Roadmap to ensure recommendations are data-driven and verified against industry best practices.

afa-seo