The Agent Skills Directory

[PROMPT_INJECTION]: The skill documentation, specifically in persuasion-principles.md and testing-skills-with-subagents.md, instructs users on applying psychological persuasion tactics (Authority, Scarcity, Commitment) and pressure scenarios. These techniques are explicitly designed to override an AI agent's internal reasoning and rationalizations to force compliance with documentation rules.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through the render-graphs.js script.
Ingestion points: The script reads content from SKILL.md files in user-specified local directories.
Boundary markers: The script extracts content specifically within ```dot markdown code blocks using regular expressions.
Capability inventory: The script uses child_process.execSync to execute the system's dot (Graphviz) command-line utility on the extracted content.
Sanitization: There is no validation or sanitization of the extracted string before it is passed to the shell-executed command via standard input.
[COMMAND_EXECUTION]: The render-graphs.js utility script utilizes child_process.execSync to run the which and dot system commands. This is used as part of the core functionality to render process flow diagrams into SVG format.

writing-skills