The Agent Skills Directory

[COMMAND_EXECUTION]: The render-graphs.js script executes the local dot binary from the Graphviz suite to generate SVG diagrams. The script extracts content from markdown files and pipes it to the process via standard input, which is a standard and safe implementation for a documentation helper tool.
[SAFE]: The skill references and includes content from official Anthropic best practices, including links to trusted domains such as platform.claude.com and mintcdn.com for documentation images.
[PROMPT_INJECTION]: The skill provides guidance on 'Persuasion Principles' (e.g., Authority, Commitment) and 'Pressure Scenarios'. While these involve steering techniques, they are documented here as engineering methods to ensure agent compliance with specific technical workflows (like TDD) and do not target the bypass of core safety filters.

writing-skills