Skip to content

Layer 2 · Signal Generators ​

Trading / allocation signals driven by composite indicators — Bull/Bear Agents, BLCV equity scores, momentum/mean-reversion strategies

Modules: 20 total · 13 Active (shipped)

Auto-generated

This page is auto-generated from card frontmatter by scripts/build-docs-layer-index.py. Do not edit by hand — changes will be overwritten. To modify the listing, edit the individual operator card or the label mapping in the script.

✅ Active (shipped) ​

Goldened, validated, and deployed to Canvas

Agent — Generic Position-Anchored Single Agent (Tier 5) ​

Canvas generic single-agent node (Society S0). One position-anchored agent (office + occupant) carries the actor model (position powers/constraints/purpose +...

Bear Agent — CASA Bearish Agent ​

Canvas Bear Agent node — symmetric to bull_agent_v1: shares the same two Canvas calls (trader-select + dual-agent-simulate), but the UI extracts `bear_...

BLCV Composite Equity Score ​

Weighted aggregation of the Business / Leadership / Competitive / Valuation dimension scores into a 0-100 composite plus a buy/sell signal.

Bull Agent — CASA Bullish Agent ​

Canvas Bull Agent node — two-step call: (1) POST /casa/trader-select computes bull/bear weights from regime plus optional macro signals (w_current / cli_dire...

CASA Strategy — LLM Strategy Code Generation (Tier 5 Attestation) ​

Mechanism DEMO / teaching vessel (repositioned 2026-08-03, YG): no academic anchor; hand-set regime weights; not an analysis tool — see References. Canva...

Engle-Granger Cointegration Test ​

Engle-Granger two-step cointegration test + hedge-ratio extraction: do two price / level series share a long-run stationary relationship; returns t-stat, p-v...

Extreme Value Theory Tail VaR / ES (POT with GPD) ​

Peaks-Over-Threshold (POT) estimation of tail VaR and Expected Shortfall (ES): convert returns to losses, take the 90th-percentile threshold, fit a GPD to th...

Fama-French 4-Factor Exposure (with Momentum) ​

OLS regression of an asset's excess return on the Fama-French 3 factors + Carhart momentum, estimating α, β_{mkt,smb,hml,mom} and R². Degrades to single-fact...

Gold Price Direction Signal (Rule + Optional XGBoost) ​

Direction signal for COMEX gold (GC=F) and other precious metals: the rule path compares current price vs SMA(50) vs SMA(200) three ways → BULLISH / BEARISH ...

Mean Reversion Strategy (Backtrader SMA deviation + real-data backtest) ​

Single-name mean-reversion backtest: Backtrader SMA(fast_period) with a ±2% deviation threshold triggering buys (deviation < -2%) / sells (deviation > +2%). ...

Momentum Strategy (Backtrader ROC + real-data backtest) ​

Single-name momentum backtest: the Backtrader ROC(period=fast_period) indicator triggers buys (ROC > 0 while flat) / sells (ROC < 0 while long). Real yfinanc...

Style / Sector Score ​

Cross-sectional ranking of a style/sector ETF basket vs a benchmark. Each name's score = RRG quadrant bonus + cross-sectional z-score of interval return and ...

Technical Analysis — TA Umbrella Operator ​

Canvas technical-analysis umbrella operator: a single call returns RSI + MACD + Bollinger + SMA/EMA/ATR main technical indicators. Each indicator's rigorous ...

📝 Draft ​

Placeholder stub — pending research references, formulas, golden test

Cascade Lens — era-by-era chain fragility ​

Assembles a set of macro time series into a directed transmission chain and reads it era by era with the cascade vocabulary: rho per era, a highlighted loop'...

Commodity Basket — a declared aggregate ​

Builds a DECLARED basket from member series with stated weights — level and return paths plus the member table. Deliberately estimation-free: a transparent o...

Conditional Amplification — response by state ​

Asks whether an outcome series responds MORE when a declared state variable is active: splits the outcome by state buckets and reports per-bucket statistics ...

Correlation Regime — driver-conditional structure ​

Asks which pairwise correlations in a declared set are DRIVER-CONDITIONAL: splits the sample by a driver state (default ENSO) and compares per-state correlat...

Driver Channel Scan — where does it bite? ​

Scans a set of DECLARED candidate channels against a driver: per-channel lead-lag profile, best lead, gain and stability, ranked. Answers where the driver ac...

Reads a DECLARED macro transmission chain link by link: per-link lead and gain from lead-lag scans, and the end-to-end read of how far and how fast a unit mo...

Value Composite — the four-dimension verdict ​

Aggregates the value suite's four dimension scores — intrinsic value, moat, quality, stewardship — under declared weights into one composite, with universe p...

Verifiable intelligence for the decisions that demand scrutiny.