The Agent Skills Directory

[SAFE]: The skill is a meta-utility for developers to create, document, and verify other skills. It provides best practices for context efficiency and discoverability.
[COMMAND_EXECUTION]: Contains a utility script render-graphs.js that executes the dot command (part of the Graphviz suite) to render diagrams. The execution uses stdin for content passing and includes safety checks such as a 10MB buffer limit, which is standard for local developer tools.
[EXTERNAL_DOWNLOADS]: References external documentation and assets from trusted domains including anthropic.com, agentskills.io, and mintcdn.com (Anthropic's documentation CDN). These references are informative and originate from well-known technical services.
[PROMPT_INJECTION]: Includes instructions on using 'Persuasion Principles' (Authority, Commitment, etc.) to design skills. These techniques are used constructively to reinforce discipline (e.g., ensuring an agent follows TDD) rather than as a means to bypass system-level safety filters or ethical guardrails.

writing-skills