The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted data from local files (existing blog posts) and web search results.
Ingestion points: Commands like /blog analyze and /blog rewrite ingest content from local files or URLs.
Boundary markers: The orchestration logic does not explicitly use delimiters or instructions to ignore embedded prompts within processed content.
Capability inventory: The workflow involves multiple agents (researcher, writer, seo, reviewer) with capabilities including file system writes, web search, and Google API interactions.
Sanitization: There is no mention of sanitizing or escaping the content retrieved from external sources before it is processed by the writing agent.
[COMMAND_EXECUTION]: The skill documentation provides examples of shell commands (e.g., using curl for bot simulation) to assist users in verifying technical SEO. These are provided as informative examples for manual execution.
[EXTERNAL_DOWNLOADS]: The skill references numerous external resources for documentation, image sourcing (Pixabay, Unsplash, Pexels), and community engagement. All sources are well-known services or author-managed platforms, and these references are documented neutrally.

blog