Markdown lint
License MIT
Prompt-injection scan
No credentials
AI safety scan: clean — A coaching skill that rigorously questions a user preparing for high-stakes scenarios (interviews, vivas, defences) by following weak answers, climbing a five-rung difficulty ladder, and producing a debrief. It includes explicit ethical guardrails (no third-party assessment, stop if distressed, don't verify claims outside its competence, don't polish false claims) and contains no hidden instructions, data exfiltration, or coercive patterns.
Approved by a human reviewer before publishing.
Every version goes through the same pipeline; the badge always describes the version you download. Automated checks reduce risk, they don’t replace reading the file before you run it.