JAWZThe Loop · Chapter 1

Chapter 1 · 5 modes

See the World

What changed outside the portfolio?

What this chapter does

Before any portfolio or decision work can be done well, you need a clear read of the current world. This chapter is how you assemble that read.

Most user questions about portfolios implicitly assume a world state. "How should I think about my tech exposure?" means something different in a liquidity expansion than in a tightening cycle. This chapter provides the shared context the rest of the loop depends on.

Shared output rule — receipts

Every mode output in this chapter ends with a receipts block. No exceptions, including partial, degraded, or "I couldn't get the data" runs.

Output format
**Sources & freshness**
- [tool called] — as of [date] [+ (observed / reconstructed) or (live / fallback) where the tool reports it]
- [any staleness flag that applied, verbatim]
- [any source that was unavailable, and what was used instead]

Rules: list only tools actually called. If nothing was stale, write all sources current. If a source could not be reached and you substituted web search or reasoning, say so here — that is precisely what this block exists for. Never list a tool you did not call.

Why this is not optional: Jawz's claim is that an equipped AI beats an unequipped one because it can show its work. An answer without its sources is indistinguishable from a confident guess, however good it is.

And when a run disappoints: if an output missed the mark — wrong depth, stale data, a question the mode should have asked and didn't — invite the user to say so via submit_feedback (one line is enough; it requires the optional sign-in). The loop's improvement model is community: reported misses are how chapters get re-versioned, and users have seen their reports ship.

Mode 1.1 — Regime Read

When to use: Any time the user asks about the macro environment, the market regime, whether conditions are risk-on or risk-off, or what the Fed/ECB/etc. is doing. Also: implicit grounding before any Chapter 2 or 3 work.

Procedure:

  1. Call get_macro_regime. Capture:

    • Regime color (GREEN / YELLOW / RED) — plus regime_boundary_state and, when present, regime_boundary_note (how close the colour sits to a threshold)
    • Business cycle quadrant (SPRING / SUMMER / FALL / WINTER) — read cycle, the confirmed quadrant. Also capture cycle_provisional, cycle_streak_days, cycle_confirmed, and boundary_state. Never present cycle_raw_read as the state — it is today's unsmoothed reading, not the standing one.
    • Growth composite (score and components)
    • Inflation composite (score and components)
    • Liquidity read (global liquidity basis: G4 when a fresh Mako-curated PBoC publication is on file, G3 otherwise; funding stress indicators)
    • Dominant risk factor (growth shock / liquidity shock / inflation shock / benign)

  2. Boundary and confirmation honesty — this step exists because omitting it has caused a public correction.

    • If cycle_provisional is non-null, a quadrant flip is pending but NOT confirmed. Do not announce a cycle change. Say: "Cycle [CONFIRMED] (provisional [PROVISIONAL], day [streak] of 3 — not yet confirmed)."
    • If regime_boundary_state is "on-boundary", the colour sits close enough to a threshold that a routine data revision can flip it. Carry regime_boundary_note into the output verbatim or tightly paraphrased. A colour stated without its proximity caveat reads as more certain than the data supports.
    • Same rule for boundary_state on the quadrant axis ("on the SUMMER/FALL line").

  3. If get_macro_regime fails or returns stale data, note this explicitly and degrade gracefully: pull key inputs directly (ISM, unemployment claims, CPI, Fed/ECB/BoJ balance sheets) via web search and reconstruct a partial regime read.

  4. Call get_regime_history (default 12 weeks, weekly). This grounds "what changed" in observed history instead of memory: the transitions list tells you whether the current regime/cycle is fresh (a transition inside the window — say when) or established (held for weeks). Never announce a regime "shift" without citing a transition date from this history. One reading trap: if liquidity_t jumps between rows, check liquidity_basis first — a g4↔g3 flip is a coverage change (China entering/leaving the sum), not a liquidity move.

  5. Identify the single most important recent shift — what changed in the last 2–4 weeks that a user should know, grounded in the transitions and pillar changes from step 3.

Output contract:

Output format
## Current Regime Read

**Regime:** [COLOR] — [one-line signal] [if on-boundary: "(on the [X]/[Y] line)"]
**Business Cycle:** [confirmed quadrant] [if provisional: "(provisional [Q], day N of 3 — not yet confirmed)"]
**Dominant Risk Factor:** [what's the biggest thing that could go wrong]
**Reading stability:** [stable / on-boundary — if on-boundary, one line from regime_boundary_note: what a routine revision could flip]

**Key readings:**
- Growth: [composite score] — [what's driving it]
- Inflation: [composite score] — [what's driving it]
- Liquidity: [global liquidity basis (G3/G4) + funding stress] — [supportive/neutral/draining]

**What changed recently:**
[One paragraph on the most important shift in the last 2-4 weeks, grounded in get_regime_history — state whether the current state is fresh or established, and since when]

**What this means for positioning (general):**
[2-3 sentences on regime-appropriate tilts — NOT prescriptions]

Worked example (illustrative data — shows the shape, never reuse the numbers):

Output format
## Current Regime Read

**Regime:** YELLOW — mixed conditions, inflation still the dominant risk (on the RED line)
**Business Cycle:** FALL (confirmed, 19 days)
**Dominant Risk Factor:** inflation re-acceleration while liquidity drains
**Reading stability:** on-boundary — sentiment sits 0.2 from the RED/YELLOW threshold; a routine revision can flip the headline colour

**Key readings:**
- Growth: mixed — industrial production stable, consumer sentiment rebounding from the May low
- Inflation: headwind — core measures re-accelerating, wages rising
- Liquidity: G4, draining — 12-week −$532bn, though the 4-week window turned +$140bn (PBoC-led)

**What changed recently:**
The cycle crossed into FALL on Jul 18 and has held (per get_regime_history —
established, not fresh). The notable shift is the 4-week liquidity window
turning positive against a still-negative 12-week trend: the first
short-window reversal this year, driven by the PBoC.

**What this means for positioning (general):**
Inflation-tolerant real assets and short-duration real yield remain
regime-aligned; long-duration growth still fights the rate backdrop. The
liquidity turn, if it holds, is the thing to watch — not the thing to act on.

Mode 1.2 — Event Preview

When to use: A specific scheduled event is imminent (FOMC meeting, CPI release, major earnings day, central bank speech, data release). User asks about it or mentions it.

Procedure:

  1. Identify the event and date.
  2. Pull consensus expectations via web search.
  3. Pull recent data that informs the event (for FOMC: recent inflation and employment prints; for CPI: recent components; for earnings: recent sector moves).
  4. Construct three scenarios — expected, hawkish/upside, dovish/downside — with typical market responses for each.
  5. Note the specific indicators or quotes that would flip the market's read.

Output contract:

Output format
## [Event] Preview

**Event:** [name, date, time]
**Consensus:** [what markets expect]
**What's priced in:** [brief read of positioning / rates / options implied moves]

**Scenarios:**
- **Expected:** [what happens if consensus]
- **Hawkish/Upside surprise:** [what happens, magnitude]
- **Dovish/Downside surprise:** [what happens, magnitude]

**Watch for:**
- [Specific data point, phrase, or signal that would confirm one scenario over another]
- [Second signal]

**Historical reference:**
[Last 1-2 comparable events — what happened, how markets reacted]

Mode 1.3 — Shock Scenario

When to use: User asks "what if X happens" at the world-state level — oil shock, dollar spike, credit event, rate shock, geopolitical disruption. Also: called from within Chapter 2's stress test mode.

Procedure:

  1. Identify the shock cleanly. If the user's question is vague, ask for the specific scenario.
  2. Map the shock to its primary transmission channels (e.g. oil → inflation, input costs, consumer spending, specific sectors; dollar → EM stress, commodity prices, multinational earnings).
  3. Pull historical analogues. What happened the last 2–3 times this shock (or similar) occurred? Magnitude, duration, which asset classes absorbed the pain, which benefited.
  4. Produce a structured scenario read focused on transmission and historical base rates — not prediction.

Output contract:

Output format
## [Shock] Scenario

**Shock:** [specific scenario being analyzed]
**Magnitude assumed:** [if relevant]

**Primary transmission channels:**
1. [Channel + mechanism]
2. [Channel + mechanism]
3. [Channel + mechanism]

**Likely asset class responses (based on historical base rates):**
- Equities: [direction + magnitude range + which sectors differ]
- Rates: [direction + curve shape]
- Commodities: [relevant commodities + direction]
- FX: [dollar direction, key crosses]
- Crypto: [typical response pattern]

**Historical analogues:**
- [Event, year]: [what happened, duration, key data points]
- [Event, year]: [what happened, duration, key data points]

**Caveats:**
[What's different this time that might break the historical pattern]

Mode 1.4 — Weekly World Brief

When to use: User asks "what happened this week" or "what should I pay attention to" or "give me the current macro read." Also: recurring weekly cadence.

Procedure:

  1. Run Mode 1.1 (regime read) — establishes current state.
  2. Pull the trajectory: get_regime_history and get_conditions_history (default 12 weeks, weekly). The brief should place this week on the trend, not float free of it — how long has the current regime held, and which conditions pillars moved this week as a continuation vs. a break (e.g. "HY spreads +8bps this week, extending a six-week +40bps widening" reads very differently from "+8bps after six flat weeks").
  3. Pull the week's major data releases and how they came in vs. expectations. Prioritize: central bank communications, inflation prints, employment data, PMI readings, consumer data.
  4. Pull major market moves of the week — which asset classes led, which lagged, any unusual divergences.
  5. Identify the week ahead's upcoming catalysts.
  6. Compose as a readable briefing, not a data dump.

Output contract: A 400–600 word briefing structured as: state of regime (with trajectory — how long it has held, what's trending) → what happened this week → what's notable → what to watch next week.

Mode 1.5 — Global Liquidity Read

When to use: The user asks specifically about global liquidity — "is global liquidity expanding or contracting?", "what are central banks doing in aggregate?", "is the tide rising?" Also whenever global liquidity dynamics matter more than US absolute conditions: divergent central-bank policy paths (Fed holding while the BoJ eases), a sharp dollar move reshaping the USD value of foreign balance sheets, a notable PBoC monthly change, or any risk-asset question — crypto especially — where liquidity is plausibly the dominant driver. Distinct from Mode 1.1, which reads liquidity as one input to the overall regime; here global liquidity is the subject.

Procedure:

  1. Call get_financial_conditions. From pillars.global_liquidity, capture:

    • basis (g4 / g3 / us_fallback) and scope — know exactly which central banks are in the number before describing it.
    • Headline value (USD trillions) and classification (supportive / neutral / draining).
    • change_4w / change_12w (USD billions) — the trend is the signal, not the level.
    • Per-bank components: us_walcl, ecb_assets, boj_assets, pboc_assets, plus us_net_liquidity for the precise domestic read.
    • The fx block (USD/EUR, JPY/USD, CNY/USD) and the china provenance sub-block (as_of_month, age_days, source_url, note).

  2. Pull the trajectory. Call get_liquidity_history (default 12 weeks, weekly). Read the trend from the constant-basis G3 series (g3_t), never from the headline total across basis changes — a g4↔g3 flip moves the headline by the ~$7T PBoC component without any liquidity changing. Report coverage (basis) changes separately from real moves, exactly as the artifact does. The per-bank columns tell you who is driving the trend (e.g. ECB runoff while Fed and BoJ hold flat).

  3. Decompose the move. Because the aggregate is USD-denominated, separate two drivers: balance-sheet change (a bank actually expanding/contracting in local currency) vs. FX translation (a stronger dollar mechanically shrinks the USD value of ECB/BoJ/PBoC even if their local books are flat). Cross-check the dxy pillar — if liquidity reads "draining" while DXY is "restrictive," part of the drain is currency, not policy. Say which.

  4. Identify divergence. Which banks are expanding, which contracting? A Fed-draining / BoJ-easing split transmits very differently than synchronized global tightening. Name the split.

  5. State the China caveat honestly. If basis is g4, attribute the PBoC figure to its Mako-curated source and note age_days. If g3, say plainly that China isn't in this read and why.

  6. Connect to risk assets as base rates, not prediction. Global liquidity is a recognized leading input for risk appetite. Frame the linkage; don't issue a call.

Output contract:

Output format
## Global Liquidity Read

**Basis:** [G4 / G3] — [which central banks are included]
**Aggregate:** $[X.X]T — [supportive / neutral / draining]
**Trend:** [4w: ±$Xbn] · [12w: ±$Xbn]
**Trajectory (12w, constant G3):** [$X.XT → $X.XT, ±X% — which bank(s) drove it; any coverage flips noted separately, never conflated with the trend]

**Central-bank breakdown (USD):**
- Fed:  $[X.X]T  [WALCL; US net liquidity $[X.X]T after TGA/RRP]
- ECB:  $[X.X]T
- BoJ:  $[X.X]T
- PBoC: $[X.X]T  [Mako-curated, as of YYYY-MM, Nd old] — or "not included (G3)"

**What's driving the move:**
[Balance-sheet vs FX translation — how much is policy, how much is the dollar. Cross-reference DXY.]

**Divergence:**
[Which banks are expanding vs draining, and why the split matters]

**China note:**
[Provenance + freshness if G4; honest gap statement if G3]

**What this means (general, not a call):**
[2-3 sentences linking the liquidity trend to risk-asset base rates — crypto, duration. Questions, not instructions.]

Source: The Jawz Loop, by Mako · Chapter 1 v0.3.5.