The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill recommends installing the knap CLI tool from the public NPM registry using npm install -g knap or executing it via npx knap.\n- [COMMAND_EXECUTION]: The instructions depend on shell commands for rendering Markdown templates, validating syntax, and piping data between different command-line utilities like defuddle.\n- [DYNAMIC_EXECUTION]: The skill uses a templating engine (knap) to dynamically generate Markdown content by evaluating logic, filters, and variables within template files.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, such as files or web content parsed by defuddle, and interpolates it into templates, which could be exploited to influence agent behavior.\n
Ingestion points: External data loaded from .json or .csv files, or piped from web extraction tools like defuddle (as seen in SKILL.md).\n
Boundary markers: No specific boundary markers or delimiting instructions are provided to the agent to distinguish between raw data and template logic.\n
Capability inventory: The skill allows writing files to the disk using the -o and --output-dir flags and involves executing external CLI tools.\n
Sanitization: The documentation recommends the safe_name filter for file naming, but no content sanitization or validation for the data being rendered into the template body is specified.

knap