The Agent Skills Directory

[SAFE]: The skill is primarily instructional, providing best practices, markdown patterns, and strategic guidance for optimizing web content for discovery by AI agents and search engines. It utilizes standard project context files (e.g., .agents/product-marketing.md) which is a common practice for AI agent configuration.
[PROMPT_INJECTION]: The skill defines a potential indirect prompt injection surface by instructing the agent to ingest external data from project-level context files.
Ingestion points: Files like .agents/product-marketing.md and .claude/product-marketing.md in the user's workspace.
Boundary markers: None explicitly defined in the instructions for these file reads.
Capability inventory: The skill itself contains no active tools but relies on the host agent's ability to read local files and potentially browse the web to perform audits.
Sanitization: None specified for the ingested content.
Risk assessment: This is categorized as a low-risk architectural choice necessary for the skill's purpose (gathering context) rather than an active exploit attempt.
[EXTERNAL_DOWNLOADS]: The skill references several external resources, including official Google Cloud documentation, tools like Semrush and Ahrefs, and the Open Knowledge Format (OKF) specification on GitHub. All referenced domains are well-known, reputable, and relevant to the skill's domain of SEO and web standards.

ai-seo