The Agent Skills Directory

[SAFE]: The skill provides a functional and documented workflow for creating agent skills. It focuses on legitimate instruction engineering and does not contain malicious code, credential exposure, or data exfiltration patterns.
[PROMPT_INJECTION]: The skill accepts user input to generate instructions as its core functionality. While this is an ingestion surface for untrusted data, the risk is mitigated as this behavior is the intended primary purpose and is governed by a multi-step workflow including user review. Ingestion points: user intent and requirements for skill creation. Boundary markers: absent. Capability inventory: uses the write_skill tool to create or update files. Sanitization: absent.

skill-creator