# GCS mission — money flow, deals, and three proofs

Settled intent: 2026-07-11  
Parent: `conversations/03-product-charter.md`  
Governing rails: GCS-SYS-001, GCS-SIG-003, GCS-DATA-002, GCS-SEC-001

## 1. One-sentence mission

**GCS tracks dollars from sovereign borrowing and central-bank balance-sheet
expansion through banks and foundations into the Financial, Military, and
Technology Industrial Complexes (FIC / MIC / TIC), focusing on political,
military, financial, and technical deals and contracts that redistribute value
and control — including technique patterns historically associated with
pump-and-dump sequences — and produces proof of intent (from the record),
proof of actions, and proof of the near future.**

### 1.1 Root path (binding product lineage)

```text
PRINT / GLOBAL $ DEBT
    → banks & primary dealers (absorb, clear, lend)
    → foundations & endowments + FIC managers (warehouse claims, vote, fees)
    → MIC · TIC · FIC operating nodes
    → redistribution CONTRACTS (awards, facilities, export licenses, equity marks)
    → return to claim-holders (interest, fees, buybacks, markups)
```

Every high-value deal in the fixture pack must answer: **where did the
borrowed or printed dollar enter, and which contract exited it?**

## 2. What we track (object)

| Track | Unit | Examples (public record only) |
| --- | --- | --- |
| **Money flow** | currency, notional, AUM, contract $ | TGA/Fed outflows → primes; 13F rotations; Form 4; buybacks; fees |
| **Capability flow** | capacity share, sole-source nodes | ASML tools, TSMC nodes, engine hot sections, cloud/AI compute |
| **Control flow** | gating share, vote power, export levers | Big Three ownership, Basel rules, sanctions, export controls |
| **Deals & transactions** | discrete events with counterparties + timestamps | USASpending awards, M&A, 13D crossings, proxy fights, armaments offsets, chip export licenses, large secondary offerings |

Complexes are **stacked, not parallel**. The same dollar may appear on three
ledgers; GCS counts **lineage and union (φ)**, never naive sums.

Domains of deals (tags, not separate systems):

- **Political** — policy, sanctions, export control, procurement priorities, Basel/IMF
- **Military** — prime contracts, co-production, FMS/DSCA where public, specialty materials
- **Financial** — issuance, ownership blocks, engagement, fees, capital structure events
- **Technical** — lithography/fab, compute, cloud, dual-use IP and standards

## 3. Technique focus: redistribution and pump–dump *signatures*

GCS does **not** open with a courtroom verdict. It opens with **technique
signatures** demonstrated from the public record. “Pump and dump” here is a
**pattern class**, not an automatic legal label:

| Phase | Observable signature (examples) | Primary bearings |
| --- | --- | --- |
| **Accumulate** | rising position, 13D/13G, Form 4 clustering, quiet block builds | 13F, 13D/G, Form 4 |
| **Inflate / catalyze** | narrative + policy + contract + media timing; engagement before vote | 8-K, DEF 14A, USASpending, news-as-clock only |
| **Extract / redistribute** | distribution, secondary, award to related nodes, fee extraction, buyback window, vote outcome moved off-record | Form 4 sells, 13F outflows, awards, N-PX residual |
| **Conceal / lock** | switching cost, sole source, regulatory moat, engagement residual | chokepoint test, residual r, export rules |

**Redistribution** is the continuous object: value, capability, and control
moving toward holders of gates (SYS-001 §2). Technique signatures are how
specific deal sequences *focus* that redistribution.

## 4. The three proofs (product spine)

This is the user-facing spine. Map every UI surface and every pipeline stage
to one of these proofs.

### 4.1 Proof of **intent** (from the record — not mind-reading)

Private mental states are **out of scope** (SYS-001 §5, §7). Intent is proven
as **declared, structural, or patterned intent**:

| Intent type | What counts as proof | What does *not* count |
| --- | --- | --- |
| **Declared** | 13D (control intent) vs 13G (passive); stated engagement policies; proxy statement strategy language; official policy texts | fabricated quotes, anonymous “they planned to…” |
| **Structural** | position that *determines* outcomes (ownership × vote × gate) — determination is a property of the position | “they wanted to hurt X” without bearings |
| **Patterned** | repeated multi-deal sequence (accumulate → catalyze → extract) with concordance across independent clocks | single chart correlation, one source |

**UI promise:** every intent card shows the filing language or structural
predicate that carries the intent claim, plus uncertainty.

### 4.2 Proof of **actions**

Actions are **events that moved money, capability, or control**, timestamped
and linked:

- Who paid whom (or who obligated whom)
- What instrument (equity, contract, debt, license, vote, export permission)
- Which complex edges the flow crossed
- Which bearings independently confirm the same transfer

**UI promise:** a deal ledger and a lineage graph; click any edge → primary
source links + as-of timestamps.

### 4.3 Proof of the **future**

Not prophecy. **Bounded, perishable forecasts** (SYS-001 §6, SIG-003):

- Next likely redistribution windows (record dates, FOMC, award cycles, proxy season)
- Probable opposition / vote outcomes under engagement intensity
- Likely positioning rotations after monetary/fiscal shocks
- Technique-phase forecasts (“accumulation late-stage → elevated extraction risk within lag L”)

Every future card shows: **point estimate · confidence band · decay / half-life ·
out-of-sample status**. In-sample-only fits never ship as trusted.

## 5. End-to-end system shape

```text
Public sources (multi-clock)
    → capture & normalize (point-in-time, authority tier)
    → entity resolution (same issuer / fund family / award recipient)
    → deal & flow graph (FIC ⇄ MIC ⇄ TIC)
    → technique signature detectors (redistribution / pump–dump phases)
    → three proof surfaces:
         INTENT  |  ACTIONS  |  FUTURE
    → immersive instrument (build/) + study narrative
```

**Product surfaces map 1:1 to proofs:**

| Instrument mode | Proof |
| --- | --- |
| Map / topology | structural intent + present control |
| Lineage | actions (money path) |
| Math lab | concentration / φ / residual magnitudes |
| Next move | future (decay mandatory) |
| Deal ledger (to build) | actions + declared intent attachments |
| Technique strip (to build) | pattern-phase of redistribution / pump–dump signatures |

## 6. Minimum viable proof stack (build order)

1. **Debt-rooted action ledger** — highest-value deals in `assets/deal_fixtures.js` (Treasury stock, QE, RRP, dealers, foreign official, foundations, SMCCF, F-35, Columbia, Ukraine supplementals, CHIPS, export controls, Nvidia/AI, common ownership, interest expense, pensions).
2. **Lineage on those deals** — dollar path drawn on the topology (print → banks → foundations → complexes → contracts).
3. **Intent attachments** — declared / structural / patterned cards per deal.
4. **Technique phase tags** — accumulate / catalyze / extract / lock with signature strength.
5. **Future strip** — deal-linked half-life + CI language, fixture-backed.
6. **Live multi-clock sources (expanded)** — browser binds in `assets/live_public_sources.js`:
   FiscalData debt + MTS net interest + auction primary-dealer take-up; FRED WALCL + RRPONTSYD;
   USASpending F-35/CHIPS keyword tempo; EDGAR 13F LMT sleeves (stamped extracts / optional live);
   **GAO F-35 lifecycle stamped** from GAO-24-106703 (sust ~$1.58T, total >$2T through 2088).
   Dual heroes: F-35 debt→lifecycle→13F; CHIPS debt→awards. App/ live clients remain phase-gated.

## 7. Hard refusals (keep the product undismissable)

- No leaked/hacked data; no personal surveillance (SEC-001 / SYS-001).
- No manufactured private transcripts of engagement meetings.
- No labeling a person or firm “pump and dump criminal” from a single residual;
  label **technique signature strength** and let the record carry the case.
- No summing FIC+MIC+TIC into a fake total GDP capture.
- No future claim without decay clause.

## 8. Success criterion

A reader can open GCS and, for a named deal sequence, answer without scrolling
to contracts:

1. **What did they intend (from the record)?**
2. **What did they do (money / capability / control moves)?**
3. **What is likely next, and when does that estimate expire?**

If any answer cannot show its bearings, it is not ready to ship.

### 8.1 Short-horizon flow future (binding product requirement)

For a **1–2 week** horizon the product must also answer:

4. **What plumbing regime are we in now?**
5. **Where does probability mass move over 7 and 14 days?**
6. **What is the entropy / decay of that forecast?**

Implemented as `assets/flow_regime.js` (Markov v1, HMM-ready) and the instrument
**Regime** mode. Predictions carry CI-via-entropy, decay half-life, and a regime
tag per GCS-SIG-003.

---

*End mission note.*