The Agent Skills Directory

[SAFE]: The skill is purely informational, consisting of documentation and implementation guides in Markdown format. It contains no executable scripts, binaries, or automated code execution patterns.
[SAFE]: All external resource references are to established and verifiable domains, including official framework documentation (dspy.ai), academic research (arxiv.org), and public code repositories (github.com/gepa-ai).
[SAFE]: The instructions incorporate proactive security and privacy advice, specifically instructing the user to process regulated data (PHI/PII) within secure boundaries using self-hosted models or specific vendor agreements.
[SAFE]: The skill advocates for robust operational practices, such as human-in-the-loop review for automatically generated instructions and the use of version control for optimized prompt artifacts.
[NO_CODE]: No functional code is provided within the skill files; it serves entirely as a reference guide for developers implementing these frameworks independently.

dspy