The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided data, creating a potential surface for indirect prompt injection.
Ingestion points: The scripts/analyze_blog.py script reads the contents of local files (Markdown, MDX, HTML) and directories provided by the user for quality auditing.
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" warnings when the agent reads and analyzes the content of these external files.
Capability inventory: The skill has the capability to read local files, write content to the filesystem via the blog-writer agent, and perform web searches via the blog-researcher agent.
Sanitization: There is no evidence of sanitization, escaping, or filtering of the ingested file content before it is processed by the agent or the analysis script.
[EXTERNAL_DOWNLOADS]: The scripts/analyze_blog.py script identifies and suggests the installation of well-known third-party libraries.
Evidence: The script mentions textstat and beautifulsoup4 as optional dependencies for advanced readability and HTML parsing metrics. These are widely recognized and established packages in the Python ecosystem.
[COMMAND_EXECUTION]: The skill documentation includes examples of command-line operations for environment testing.
Evidence: The references/ai-crawler-guide.md file contains curl and bash examples to help users or developers verify AI crawler visibility and server-side rendering. These are provided as educational templates and are not executed automatically by the agent's instructions.

blog