The Agent Skills Directory

[COMMAND_EXECUTION]: The provided render-graphs.js utility script uses child_process.execSync to invoke the Graphviz dot command. This is a legitimate developer tool used to convert diagram source code within SKILL.md into SVG images. The script passes content to the external process via standard input, which is a secure method for process communication.
[DYNAMIC_EXECUTION]: The render-graphs.js script dynamically extracts and processes DOT language blocks from the skill's markdown documentation at runtime to generate visualization files.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill's core protocols explicitly require agents to redact secrets and personally identifiable information (PII) before sharing context with subagents, following the principle of least privilege.
[PROMPT_INJECTION]: The skill includes advanced instructions on 'bulletproofing' prompts to resist rationalization and bypass attempts. While these involve psychological persuasion principles, they are utilized here as a safety mechanism to ensure the agent adheres to rigorous development standards (like TDD) rather than as a malicious injection.

writing-skills