Hallucination Scanner
Runs 15 detection rules against any AI-generated response — identifying fabricated citations, precision abuse, instruction drift, absolute certainty language, and logical inconsistency. Outputs a structured Claim → Evidence → Flaw evaluation exportable as a timestamped .txt for annotation logs and QA pipelines. Rule 6 (instruction compliance) activates when the original prompt is supplied.
Negotiation Trainer
Interactive practice with a tone-reactive AI counterpart. Push too hard and they push back; be collaborative and they open up. Five scenarios: Salary, Business Deal, Vendor, Conflict Resolution, Everyday. Real-time style feedback and confidence scoring. Strengths and improvement suggestions after each exchange. Runs entirely in the browser — no API calls.
Prompt Router
Paste a task description. The router scores it across 6 dimensions and suggests a starting model based on the CYW VLAN architecture — Perplexity through ChatGPT. Rationale first, recommendation second. Use it as an aid, not a command.
Interactive Shared Responsibility Model
A filterable compliance matrix mapping Provider and Customer responsibilities across IaaS, PaaS, and SaaS environments. Overlay compliance frameworks — HIPAA, GDPR, SOC 2, ISO 27001 — to see exactly where provider responsibility ends and customer responsibility begins. Expandable categories, PDF export, print-ready, dark mode.
More tools in development.
Built by Contruil LLC as part of the human-governed AI orchestration stack.