The Agent Skills Directory

[SAFE]: The skill serves as a comprehensive educational resource for the instructor library, which is a widely recognized open-source project for structured LLM interactions.
[SAFE]: Code examples utilize safe placeholders for sensitive information such as API keys (your-api-key) and local service endpoints (ollama), which is a best practice for security documentation.
[SAFE]: The skill directs users to the official project documentation (python.useinstructor.com) and the authoritative public repository on GitHub.
[INDIRECT_PROMPT_INJECTION]: The skill demonstrates processing external text for data extraction (e.g., 'CompanyInfo' and 'SentimentAnalysis' examples in SKILL.md). Boundary markers are represented by instructional text within the prompt context. The skill's primary capability is orchestrating requests to LLM provider APIs. Sanitization is provided through the library's main functionality: Pydantic schema validation and automatic retry logic when model output fails to meet structural requirements.
[DYNAMIC_EXECUTION]: The documentation illustrates the use of pydantic.create_model for runtime schema generation. This is a standard feature of the Pydantic library used for flexible data validation and does not introduce unusual security risks in this context.

instructor