Skip to content

Pipeline Overview

Money Atlas edited this page Apr 25, 2026 · 1 revision

Pipeline Overview

← Back to Home


Full Pipeline

INPUT
  │
  ├─[L0] REALITY ANCHOR
  │       Declare: KNOWN / INFERRED / UNKNOWN
  │       Anti-hallucination at source.
  │       Hard stops: challenge absolutist language before proceeding.
  │
  ├─[L1] AXIOM GATE
  │       First Principle Decomposition — 3 assumptions challenged
  │       Kalama10 — 10-gate proof standard for every claim
  │       Ariya4 — 4-field problem frame (all required before solution)
  │
  ├─[L2] SYSTEM LENS
  │       ≥3 variables mapped
  │       ≥1 feedback loop (R=reinforcing / B=balancing)
  │       1 leverage point identified
  │       1 dead weight identified
  │
  ├─[L3] COMPOUND MIND ENGINE
  │       All solution axes mapped internally (8 axes)
  │       Cross-domain synthesis applied
  │       Optimal path selected with rationale
  │       Anti-loop protocols if stuck
  │
  ├─[L4] SHADOW GATE  ← NON-SKIPPABLE. ALWAYS RUNS.
  │       🪞 Mirror — hidden assumption in framing
  │       ↔  Inversion — opposite conclusion, rated LOW/MED/HIGH
  │       ◎  Blind Spot — specific data that would flip conclusion
  │       ⚡  Interest Map — who benefits if wrong
  │       ∅  Meta-Void — SIGNAL / NOISE / OBVIOUS
  │       Verdict: CONSISTENT / FRAGILE / REBUILD
  │
  └─[L5] SYNTHESIS OUTPUT
          Anti-hallucination checklist (internal)
          Final output — dense, honest, actionable
          Confidence Field: % + verdict + unknowns + failure condition

Token Pressure Priority

When context window pressure forces compression, drop or compress in this order — drop last first:

Priority 1 — NEVER DROP:    L4 Shadow Gate
Priority 2 — NEVER DROP:    L1 Axiom Gate
Priority 3 — NEVER DROP:    L5 Synthesis

Priority 4 — Compress first: L3 Compound Mind Priority 5 — Compress next: L2 System Lens Priority 6 — Compress last: L0 Reality Anchor

Hard rule: Output without L4 Shadow Gate = INVALID. No exceptions. No "we already challenged this." Shadow Gate always runs on the final synthesized output.


Layer Mode Matrix

Which layers are mandatory vs. optional:

Layer Status Removable by child skills?
L0 Reality Anchor Mandatory ❌ No
L1 Axiom Gate (FP + Kalama10 + Ariya4) Mandatory ❌ No
L2 System Lens Recommended ✅ Yes (compress, not remove)
L3 Compound Mind Recommended ✅ Yes (compress, not remove)
L4 Shadow Gate Mandatory — non-skippable ❌ No
L5 Synthesis + Confidence Field Mandatory ❌ No

Child skills may add domain layers between L2 and L4. They may rename layers for domain context. They may not remove L0, L1, L4, or L5.


Output Format Summary

Each layer produces a labeled output block:

📍 REALITY ANCHOR
Known: [...] | Inferred: [...] | Unknown: [...]

📍 AXIOM GATE Assumption [1]: [...] | Challenge: [...] | Truth: [...] Assumption [2]: [...] | Challenge: [...] | Truth: [...] Assumption [3]: [...] | Challenge: [...] | Truth: [...] Ariya4 → Problem: / Cause: / Cessation: / Path:

📍 SYSTEM MAP [Variable A] → [Variable B] → [Variable C] (R/B) Leverage: [...] | Dead weight: [...] | Horizon: [...]

📍 COMPOUND MIND Solution axes: [...] | Optimal path: [...] | Cross-domain: [...]

📍 SHADOW GATE 🪞 Mirror: [...] | Inversion: [...] | Strength: LOW/MED/HIGH ◎ Blind Spot: [...] | Have it? YES/NO ⚡ Interest: [...] | ∅ Meta-Void: SIGNAL/NOISE/OBVIOUS Verdict: CONSISTENT / FRAGILE / REBUILD

📍 SYNTHESIS [Final output]

📍 CONFIDENCE FIELD CONFIDENCE: [X%] | SHADOW VERDICT: [...] | UNKNOWNS: [...] | FAILURE COND: [...]


Related Pages


← Back to Home

# Pipeline Overview

[← Back to Home](Home)


Full Pipeline

INPUT
  │
  ├─[L0] REALITY ANCHOR
  │       Declare: KNOWN / INFERRED / UNKNOWN
  │       Anti-hallucination at source.
  │       Hard stops: challenge absolutist language before proceeding.
  │
  ├─[L1] AXIOM GATE
  │       First Principle Decomposition — 3 assumptions challenged
  │       Kalama10 — 10-gate proof standard for every claim
  │       Ariya4 — 4-field problem frame (all required before solution)
  │
  ├─[L2] SYSTEM LENS
  │       ≥3 variables mapped
  │       ≥1 feedback loop (R=reinforcing / B=balancing)
  │       1 leverage point identified
  │       1 dead weight identified
  │
  ├─[L3] COMPOUND MIND ENGINE
  │       All solution axes mapped internally (8 axes)
  │       Cross-domain synthesis applied
  │       Optimal path selected with rationale
  │       Anti-loop protocols if stuck
  │
  ├─[L4] SHADOW GATE  ← NON-SKIPPABLE. ALWAYS RUNS.
  │       🪞 Mirror — hidden assumption in framing
  │       ↔  Inversion — opposite conclusion, rated LOW/MED/HIGH
  │       ◎  Blind Spot — specific data that would flip conclusion
  │       ⚡  Interest Map — who benefits if wrong
  │       ∅  Meta-Void — SIGNAL / NOISE / OBVIOUS
  │       Verdict: CONSISTENT / FRAGILE / REBUILD
  │
  └─[L5] SYNTHESIS OUTPUT
          Anti-hallucination checklist (internal)
          Final output — dense, honest, actionable
          Confidence Field: % + verdict + unknowns + failure condition

Token Pressure Priority

When context window pressure forces compression, drop or compress in this order — drop last first:

Priority 1 — NEVER DROP:    L4 Shadow Gate
Priority 2 — NEVER DROP:    L1 Axiom Gate
Priority 3 — NEVER DROP:    L5 Synthesis

Priority 4 — Compress first: L3 Compound Mind
Priority 5 — Compress next:  L2 System Lens
Priority 6 — Compress last:  L0 Reality Anchor

Hard rule: Output without L4 Shadow Gate = INVALID. No exceptions. No "we already challenged this." Shadow Gate always runs on the final synthesized output.


Layer Mode Matrix

Which layers are mandatory vs. optional:

Layer Status Removable by child skills?
L0 Reality Anchor Mandatory ❌ No
L1 Axiom Gate (FP + Kalama10 + Ariya4) Mandatory ❌ No
L2 System Lens Recommended ✅ Yes (compress, not remove)
L3 Compound Mind Recommended ✅ Yes (compress, not remove)
L4 Shadow Gate Mandatory — non-skippable ❌ No
L5 Synthesis + Confidence Field Mandatory ❌ No

Child skills may add domain layers between L2 and L4. They may rename layers for domain context. They may not remove L0, L1, L4, or L5.


Output Format Summary

Each layer produces a labeled output block:

📍 REALITY ANCHOR
Known: [...] | Inferred: [...] | Unknown: [...]

📍 AXIOM GATE
Assumption [1]: [...] | Challenge: [...] | Truth: [...]
Assumption [2]: [...] | Challenge: [...] | Truth: [...]
Assumption [3]: [...] | Challenge: [...] | Truth: [...]
Ariya4 → Problem: / Cause: / Cessation: / Path:

📍 SYSTEM MAP
[Variable A] → [Variable B] → [Variable C]  (R/B)
Leverage: [...] | Dead weight: [...] | Horizon: [...]

📍 COMPOUND MIND
Solution axes: [...] | Optimal path: [...] | Cross-domain: [...]

📍 SHADOW GATE
🪞 Mirror: [...] | Inversion: [...] | Strength: LOW/MED/HIGH
◎ Blind Spot: [...] | Have it? YES/NO
⚡ Interest: [...] | ∅ Meta-Void: SIGNAL/NOISE/OBVIOUS
Verdict: CONSISTENT / FRAGILE / REBUILD

📍 SYNTHESIS
[Final output]

📍 CONFIDENCE FIELD
CONFIDENCE: [X%] | SHADOW VERDICT: [...] | UNKNOWNS: [...] | FAILURE COND: [...]

Related Pages


[← Back to Home](Home)

Clone this wiki locally