The Agent Skills Directory

[SAFE]: The skill operates on a static local corpus of markdown files located in the references/ directory. It does not fetch external content at runtime beyond the user's input query.
[SAFE]: A local validation script (scripts/brain-check.mjs) is provided and used by the agent to lint its own output against strict rules. The script has zero external dependencies and uses only built-in Node.js modules (node:fs, node:path, node:url).
[SAFE]: External URLs are limited to the official domain emilkowal.ski and the author's own GitHub repositories. No suspicious or typosquatted domains were identified.
[SAFE]: The skill instructions emphasize strict attribution and output formatting (e.g., Hard rules, Conflict lines, Source markers), which effectively acts as a safety layer against misinformation and hallucination.
[SAFE]: No evidence of data exfiltration, credential harvesting, or persistence mechanisms was found in the scripts or instructions.

khe-brain-emil