The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses uv run to execute a local shim script (scripts/invoke.py) and a library script ($LIBRARY/scripts/languages.py) to perform its core functions.
[DYNAMIC_EXECUTION]: The scripts/invoke.py script is designed to discover and execute a local engine script (kntnt.py) by searching across common agent configuration directories (e.g., ~/.claude/skills). It invokes the engine using subprocess.run, passing user input through standard input.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from external sources to draft or review writing briefs.
Ingestion points: External material provided as text, file paths, or URLs (SKILL.md).
Boundary markers: User-provided material is passed directly to the processing engine via stdin.
Capability inventory: The skill has the ability to execute scripts via uv run and write output to specified delivery files in the library.
Sanitization: The skill mitigates risks by strictly mapping external content to the 13-question template, identifying inferences, and marking missing information rather than executing instructions found in the material.

brief