Skip to content

Lycheetah/cascade-cicada-tool-

Repository files navigation

CASCADE CICADA TOOL · TCCT

Desktop knowledge scoring engine. Built on the CASCADE epistemic framework.

Score any claim, argument, or body of research against a structured 9-layer framework. Get a number, not a vibe.


Download

→ Download CASCADE CICADA TOOL v0.1.0 (Windows)

Portable .exe — no install required. Run it directly.


What It Does

Most note tools help you collect. TCCT helps you prove.

There is a difference between knowing something and knowing it well enough that it holds when challenged. TCCT gives every claim a score derived from the actual depth and coherence of what you've written — not a vibe, a number.


The Scoring Framework

Score Band Meaning
1–20 WEAK Claim doesn't hold yet. Evidence thin or missing.
21–40 DEVELOPING Partial support. Structure forming.
41–60 MIDDLE Holds under moderate pressure. Respectable.
61–80 STRONG Well-evidenced, coherent, survives scrutiny.
81–100 FOUNDATION Load-bearing. Knowledge you build on.
101–999 FRONTIER Beyond current verification. Held with conviction, not yet provable.

999 is not an error. It is reserved for claims that exist ahead of the evidence.


The 9 Onion Layers

Every knowledge block is scored across 9 layers from core to edge:

Layer Role
AXIOM The irreducible core claim. If this fails, the block fails.
FOUNDATION Primary evidence. What holds the axiom up.
STRUCTURE Logical architecture connecting claim to evidence.
COHERENCE Internal consistency. Does the block contradict itself?
RESONANCE Connections to other known truths.
TENSION Where the claim meets genuine friction. This is good — tension means it's real.
CONTESTED Active dispute zone. What others challenge.
SPECULATIVE What the claim implies beyond what is proven.
FRONTIER The unknown edge. What this claim cannot yet account for.

The Truth Pressure Formula

Π = E · P / S

  • E — Evidence density (FOUNDATION + STRUCTURE average)
  • P — Power (AXIOM score — the load-bearing claim)
  • S — Effective coherence (COHERENCE compressed by TENSION and CONTESTED)

High Π + high coherence = strong stable knowledge.
High Π + low coherence = knowledge under pressure, restructuring imminent.


Three Scoring Frameworks

  • CASCADE — Does this knowledge hold under contradiction pressure?
  • AURA — Does this align with 7 constitutional invariants (human primacy, inspectability, honesty, reversibility, non-deception, memory continuity, care as structure)?
  • LAMAGUE — Can this claim be formalised without losing meaning?

Features

  • Pyramid → File → Block → 9 Onion Layer architecture
  • Dual-track scoring: AI framework score + your own sovereign score (1–999)
  • Live Π truth pressure display per block
  • CASCADE event detection (score drops flagged with severity)
  • Cross-block tension detection
  • Layer dependency enforcement (FOUNDATION ≤ AXIOM×1.1, STRUCTURE ≤ FOUNDATION×1.2)
  • AI block suggestion + scoring via DeepSeek (bring your own API key)
  • AURA 7-invariant evaluator
  • Resonance / contradiction / synthesis experiment engine
  • SVG pyramid visualisation
  • Native Windows export

Running from Source

git clone https://github.com/Lycheetah/cascade-cicada-tool-.git
cd cascade-cicada-tool-
npm install

# Dev mode (two terminals)
npm run vite
npm run electron

Requires Node.js 18+. No native dependencies.


Provenance

The scoring engine, onion structure, and truth pressure mechanics are derived from 250 pages of original research compressed into the CODEX_AURA_PRIME — 1,402 pages total, 10 frameworks, continuous development.

Built by Mackenzie Conor James Clark · Lycheetah Framework


Part of the Lycheetah Ecosystem

Sol Lite          — gateway mobile app (Play Store)
Sol (full)        — 4 personas, full framework, BYOK
TCCT              — this tool — desktop knowledge OS
CASCADE API       — post-funding: scoring engine as infrastructure

Two points. One Work. The Gold belongs to neither.

About

CASCADE CICADA TOOL (TCCT) — Desktop knowledge scoring engine

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages