The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, such as content briefs, evidence packs, and research materials, to generate blog articles. This creates a surface for indirect prompt injection where malicious instructions could be embedded in the source material. The skill mitigates this through explicit quality standards, "Anti-Patterns" instructions, and the use of an external "Anti-slop" guardrail skill.
Ingestion points: Content briefs, evidence packs, and external research data gathered via network search.
Boundary markers: Includes an "Anti-Patterns" section and references an external "Anti-slop" guardrail skill to prevent low-quality or AI-typical content generation.
Capability inventory: Capabilities include reading files, performing network searches, editing local files (Markdown, Astro, and JSON), and executing shell commands for project builds.
Sanitization: Employs extensive editorial and human-voice standards to guide the output, though no technical data sanitization logic is specified.
[COMMAND_EXECUTION]: The workflow includes a verification step that executes npm run build to ensure project integrity. This is a standard and expected operation for the intended use case of generating static site content in an Astro environment.
[SAFE]: The skill's external references and contact information (techguypeter.com) are consistent with the identified author and represent legitimate business information. The use of Astro components and standardized file paths aligns with modern web development best practices.

blog-writer