The Agent Skills Directory

[COMMAND_EXECUTION]: The script render-graphs.js serves as a utility for converting Graphviz DOT blocks in markdown files into SVG images. It invokes the system's local dot binary using child_process.execSync with content passed via standard input to prevent command injection, and it employs strict alphanumeric validation for graph names to mitigate path traversal risks during file generation.
[PROMPT_INJECTION]: The skill includes documentation (e.g., persuasion-principles.md and testing-skills-with-subagents.md) that guides the author on using authoritative framing and commitment-based instructions to improve the agent's adherence to discipline-enforcing workflows. These techniques are focused on process reliability and do not target the bypass of safety guardrails or core system prompts.
[EXTERNAL_DOWNLOADS]: Documentation within the skill contains image references pointing to mintcdn.com, the content delivery network for official Anthropic documentation. These references are used for instructional support and do not involve the execution of untrusted external code.

writing-skills