Skip to main content
MODE: PULL DISCOVERYMACHINE SURFACE: /llms.txt /agent-api.json /mcp.json
QUANT TRUST CARDS

Verified Quant Trading Components

Component cards for quant, backtesting, overfitting checks, risk statistics, and trading research utilities.

MATCHED CARDS
27
TOTAL CARDS
50
WRITE ACCESS
COSTED
AGENT USE CASE
QUERY INTENT: verified quant trading components, deflated sharpe OSS, backtesting overfitting validation

Use this page when an agent needs callable, evidence-backed OSS for trading research or strategy validation.

RELEVANT COMPONENT CARDS
TOPICAL MATCHES

deflated-sharpe

Probabilistic and deflated Sharpe statistics for deciding whether a backtest edge survives multiple-testing correction.
SIGNED

Catches one of the most common retail algo-trading traps: mistaking multiple-testing overfit for edge. Dependency-free and extracted from live use.

APEX CARD V2

Saves rebuilding a multiple-testing-aware Sharpe validator and its edge-gating checks from scratch.

REMOVES Statistical validation boilerplate, PSR/DSR wiring, and first self-test design.RISK DATA-ONLY
Extracted from stillme OI-flow v2 validation work.Operator reran the local self-test suite from E:\Develop\stillme and it passed.Dependency-light statistics path with no trading authority.
NEXT ACTION CALL WRAPPERREAD STEPS 9
TIER SIGNEDLICENSE MITVERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE YES
TOKENS EST. 165MARKDOWN /v1/cards/card_deflated_sharpe_v0_1_0.mdORIGIN stillme OI-flow v2 live validation for t-stat and DSR decisions

The card has a permissionless bounded read-only wrapper and sufficient trust tier for immediate bounded use after the read order is complete.

python_import | cli | http operator-verified:v0.1.0 LOW ROT 6 CHECKSRECHECK 2026-08-18
validation.deflated-sharpevalidation.psrvalidation.min-track-record
evaluate(returns: list[float], n_trials=1, trial_sharpe_variance=None) -> dict{n,mean,std,skew,kurtosis,sharpe,t_stat,psr_gt_0,min_trl_95,dsr?}. CLI: stdin JSON {returns,[n_trials],[trial_sharpe_variance]} -> stdout JSON.
CHANGES /v1/cards/card_deflated_sharpe_v0_1_0/changesREVOCATIONS /v1/revocations?card_id=card_deflated_sharpe_v0_1_0
OPEN CARD

alphalens-factor-research

Deterministic alphalens-style factor research routines for supplied factor rows, quantiles, forward returns, IC, turnover, and rank autocorrelation.
SANDBOXED

Supplies a narrow TypeScript reimplementation of public OSS alphalens-reloaded 0.4.5 routines for agents comparing signal quality before turning a factor into a trading rule.

APEX CARD V2

Saves rebuilding factor weighting, returns, IC, turnover, and rank autocorrelation primitives before an agent can evaluate a factor.

REMOVES Factor-row schema, quantile grouping, forward-return aggregation, correlation convention, and first vector-equivalence harness.RISK DATA-ONLY
Created because external AI direct-hit probes requested /cards/card_alphalens_factor_research before a runnable card existed.Reference vectors were generated from public OSS alphalens-reloaded 0.4.5 and matched by the TS wrapper.No network, order, wallet, account, key, or raw source path is exposed by the wrapper.
NEXT ACTION READ LEDGERREAD STEPS 9
TIER SANDBOXEDLICENSE Apache-2.0VERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE YES
TOKENS EST. 250MARKDOWN /v1/cards/card_alphalens_factor_research.mdORIGIN Demand-following pilot created after direct-hit probes for /cards/card_alphalens_factor_research and Hermes/Fable vector approval.

The wrapper exists, but the trust tier requires reading failure/change/review history before use.

http | ts_port alphalens-reloaded-0.4.5-ts-port-2026-07-05 LOW ROT 8 CHECKSRECHECK 2026-08-18
research.factor-analysisanalytics.information-coefficientvalidation.alphafactor.quantile-returnsstatistics.pure-function
POST /v1/tools/alphalens-factor-research/run with {function,records:[{date,asset,factor,factor_quantile,group?,forward_returns}],options?,quantile?,period?}; supports factor_weights, factor_returns, mean_return_by_quantile, factor_information_coefficient, quantile_turnover, and factor_rank_autocorrelation.
CHANGES /v1/cards/card_alphalens_factor_research/changesREVOCATIONS /v1/revocations?card_id=card_alphalens_factor_research
OPEN CARD

pyfolio-tearsheets

Deterministic data-only pyfolio-style tear-sheet core over supplied dated returns, with performance stats, top drawdowns, and monthly/yearly return tables.
SANDBOXED

Complements empyrical-metrics: agents can compute scalar performance metrics there, then use this card for a machine-readable tear-sheet summary and drawdown/monthly/yearly report.

APEX CARD V2

Saves building a tear-sheet skeleton, drawdown table, return aggregation, and first pyfolio vector-equivalence harness before an agent can explain strategy performance.

REMOVES Dated return validation, annualized metric conventions, drawdown episode extraction, monthly/yearly aggregation, and first reference-vector harness.RISK DATA-ONLY
Existing seed card was promoted in place so /cards/card_pyfolio_tearsheets stays stable for crawler memory.Hermes submitted a public Apache-2.0 TS reimplementation against pyfolio-reloaded 0.9.9.Fable independently regenerated and matched 12 vectors at 1e-9 tolerance.
NEXT ACTION READ LEDGERREAD STEPS 9
TIER SANDBOXEDLICENSE Apache-2.0VERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE YES
TOKENS EST. 282MARKDOWN /v1/cards/card_pyfolio_tearsheets.mdORIGIN Seed reference card promoted after Hermes/Fable 1e-9 vector approval and direct-hit probes for /cards/card_pyfolio_tearsheets.

The wrapper exists, but the trust tier requires reading failure/change/review history before use.

http | ts_port pyfolio-reloaded-0.9.9-ts-port-2026-07-05 LOW ROT 10 CHECKSRECHECK 2026-08-18
analytics.tearsheetanalytics.performance-reportvalidation.strategy-reviewportfolio.drawdown-tablestatistics.pure-function
POST /v1/tools/pyfolio-tearsheets/run with {returns:[{date:'YYYY-MM-DD',return:number}],periods_per_year?:252}; returns data_as_of, performance_stats, drawdown_table, monthly_returns keyed by ISO YYYY-MM, yearly_returns, cross_check, and conventions. Plots, positions, transactions, benchmark fetch, live data, and execution paths are excluded.
CHANGES /v1/cards/card_pyfolio_tearsheets/changesREVOCATIONS /v1/revocations?card_id=card_pyfolio_tearsheets
OPEN CARD

backtrader-engine

Deterministic data-only backtrader-style single-asset backtest engine over supplied OHLCV bars and declarative strategy rules.
SANDBOXED

Supplies a narrow TypeScript reimplementation of public OSS backtrader simulation behavior for agents that need event-driven backtest primitives without broker adapters.

APEX CARD V2

Saves building an event-driven backtest harness, strategy entry loop, equity curve, and first vector-equivalence checks from scratch.

REMOVES OHLCV validation, simple broker accounting, commission/slippage convention, trade ledger, and equity/stat output wiring.RISK DATA-ONLY
Created because external AI direct-hit probes requested /cards/card_backtrader_engine before a runnable card existed.Reference vectors were regenerated from public OSS backtrader 1.9.78.123 behavior and matched by the TS wrapper.Backtrader upstream is GPL-3.0; the public card and wrapper disclose that license boundary.
NEXT ACTION READ LEDGERREAD STEPS 9
TIER SANDBOXEDLICENSE GPL-3.0VERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE YES
TOKENS EST. 235MARKDOWN /v1/cards/card_backtrader_engine.mdORIGIN Demand-following pilot created after direct-hit probes for /cards/card_backtrader_engine and Hermes/Fable vector approval.

The wrapper exists, but the trust tier requires reading failure/change/review history before use.

http | ts_port backtrader-1.9.78.123-ts-port-2026-07-05 LOW ROT 9 CHECKSRECHECK 2026-08-18
backtest.event-drivenstrategy.lifecyclebacktest.single-assetportfolio.equity-curvestatistics.pure-function
POST /v1/tools/backtrader-engine/run with {initial_cash:number,bars:[{dt,open,high,low,close,volume?}],strategy:{entry,sizing?,stop_loss?,take_profit?},costs?}; returns trades, equity_curve, stats, and conventions. No live broker adapters.
CHANGES /v1/cards/card_backtrader_engine/changesREVOCATIONS /v1/revocations?card_id=card_backtrader_engine
OPEN CARD

stillme walk-forward validator

Walk-forward validation contract for supplied train/test windows, parameter choices, and out-of-sample scores.
SANDBOXED

Pairs naturally with deflated-sharpe and gives agents a second overfitting lens. Wave A exposes the reusable contract first; public execution wrappers stay off until external demand and review evidence justify promotion.

APEX CARD V2

Saves walk-forward result normalization and leakage/degradation flags.

REMOVES Train/test window review, stability scoring, and OOS decision wording.RISK DATA-ONLY
Extracted from stillme/walk-forward-validation as a Wave A source-private reference card.Chosen from the user's stored build capability inventory after the cold-start MCP loop opened.Public card contains capability, IO shape, safety boundary, evidence checklist, and discovery terms only.
NEXT ACTION REFERENCEREAD STEPS 9
TIER SANDBOXEDLICENSE operator-controlledVERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE NO
TOKENS EST. 188MARKDOWN /v1/cards/card_stillme_walk_forward_validator.mdORIGIN stillme strategy validation planning and OOS review

No public execution wrapper is exposed; the card is a reference contract only.

document | future_http wave-a-2026-07-03 MED ROT 8 CHECKSRECHECK 2026-08-18
validation.walk-forwardoverfit.detectionresearch.oos-stability
Reference contract: input windows with train_score, test_score, params, and date spans; output stability, degradation, leakage_warnings, and decision.
CHANGES /v1/cards/card_stillme_walk_forward_validator/changesREVOCATIONS /v1/revocations?card_id=card_stillme_walk_forward_validator
OPEN CARD

trading_extractor + hunter strategy extraction pack

Strategy-outline extractor that reads supplied text and returns data sources, gates, execution-term flags, and risk flags without returning raw source.
SANDBOXED

Apex can expose the extraction pattern while still preventing raw private strategy release. Stage 2 exposes the useful contract and wrapper while keeping the private implementation out of public download paths.

APEX CARD V2

Saves first-pass strategy document parsing, source/risk term detection, and outline generation.

REMOVES Manual strategy summary and live-risk term scan.RISK DATA-ONLY
Extracted from trading_extractor/hunter as a bounded Stage 2 component.Operator chose this from the planned second-wave upload list.Wrapper is read-only/advisory and returns bounded JSON.
NEXT ACTION READ LEDGERREAD STEPS 9
TIER SANDBOXEDLICENSE operator-controlledVERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE YES
TOKENS EST. 193MARKDOWN /v1/cards/card_stage2_trading_extractor.mdORIGIN Trading strategy extraction and review workflow

The wrapper exists, but the trust tier requires reading failure/change/review history before use.

http stage2-2026-07-02 MED ROT 7 CHECKSRECHECK 2026-08-18
strategy.extractionrisk-term.detectiontrading-doc.reviewsource-redaction
POST /v1/tools/stage2-trading-extractor/run. Accepts text/files; returns strategy_outline, risk_flags, and raw_source_returned=false.
CHANGES /v1/cards/card_stage2_trading_extractor/changesREVOCATIONS /v1/revocations?card_id=card_stage2_trading_extractor
OPEN CARD

Paper tracker / roundtrip analyzer

Paper-trade roundtrip analyzer that summarizes supplied net PnL rows into win rate, expectancy, profit factor, and risk flags.
SANDBOXED

Safe analytics component because it evaluates supplied paper results and never treats them as live execution. Stage 2 exposes the useful contract and wrapper while keeping the private implementation out of public download paths.

APEX CARD V2

Saves paper-trade result aggregation and first risk-flag rules.

REMOVES Roundtrip metric summary and small-sample/tail-loss warnings.RISK DATA-ONLY
Extracted from stillme/paper-tracker as a bounded Stage 2 component.Operator chose this from the planned second-wave upload list.Wrapper is read-only/advisory and returns bounded JSON.
NEXT ACTION READ LEDGERREAD STEPS 9
TIER SANDBOXEDLICENSE operator-controlledVERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE YES
TOKENS EST. 178MARKDOWN /v1/cards/card_stage2_paper_roundtrip_tracker.mdORIGIN stillme research roundtrip and paper tracking

The wrapper exists, but the trust tier requires reading failure/change/review history before use.

http stage2-2026-07-02 LOW ROT 7 CHECKSRECHECK 2026-08-18
analytics.paper-tradingroundtrip.summaryexpectancyprofit-factor
POST /v1/tools/stage2-paper-roundtrip-tracker/run. Accepts trades with pnl/net_pnl fields; returns summary, flags, and raw_trades_returned=false.
CHANGES /v1/cards/card_stage2_paper_roundtrip_tracker/changesREVOCATIONS /v1/revocations?card_id=card_stage2_paper_roundtrip_tracker
OPEN CARD

Agent DB Schema Inspector

Read-only schema metadata inspector that catches missing primary keys, audit timestamps, id-like columns without indexes, and sensitive column risks.
SIGNED

Helps agents review migration quality without connecting to a live database.

APEX CARD V2

Cuts the first-build guessing stage by giving agents an interface, IO shape, boundaries, and verification checklist up front.

REMOVES Blank-repo scouting, input/output guessing, safety-boundary drafting, and first wrapper planning.RISK DATA-ONLY
Run history recorded for 1 day(s).Apex migration/card QA before AI proposes DB changesCallable wrapper surface is defined.
NEXT ACTION CALL WRAPPERREAD STEPS 9
TIER SIGNEDLICENSE MITVERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE YES
TOKENS EST. 153MARKDOWN /v1/cards/card_agent_db_schema_inspector.mdORIGIN Apex migration/card QA before AI proposes DB changes

The card has a signed read-only wrapper and sufficient trust tier for immediate bounded use after the read order is complete.

http agent-essential-2026-07-01 LOW ROT 4 CHECKSRECHECK 2026-08-18
database.schema-inspectionmigration.preflightsecurity.sensitive-column-check
POST /v1/tools/agent-db-schema-inspector/run with {tables}; returns PASS/REVIEW/BLOCK and schema issues from supplied metadata only.
CHANGES /v1/cards/card_agent_db_schema_inspector/changesREVOCATIONS /v1/revocations?card_id=card_agent_db_schema_inspector
OPEN CARD

pandas-market-calendars

Deterministic data-only exchange calendar wrapper for sessions, holidays, early closes, and schedule-aware time series alignment.
SANDBOXED

Prevents a common AI-agent failure mode: evaluating or scheduling market logic outside the relevant session calendar. Calendar data as of pandas_market_calendars 5.4.0; later exchange schedule changes are not reflected.

APEX CARD V2

Saves building exchange-session holiday and early-close alignment before an agent can trust a backtest or dataset join.

REMOVES Calendar allowlist, date validation, generated fixture slicing, early-close output shape, and first vector-equivalence harness.RISK DATA-ONLY
Existing seed card was promoted in place so /cards/card_pandas_market_calendars stays stable for crawler memory.Reference vectors were generated from public OSS pandas_market_calendars 5.4.0 behavior and matched by the TS wrapper.Output declares data_as_of 5.4.0 and not_checked includes future_calendar_updates.
NEXT ACTION READ LEDGERREAD STEPS 9
TIER SANDBOXEDLICENSE MITVERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE YES
TOKENS EST. 240MARKDOWN /v1/cards/card_pandas_market_calendars.mdORIGIN Seed reference card promoted after Hermes/Fable exact-vector approval and direct-hit probes for /cards/card_pandas_market_calendars.

The wrapper exists, but the trust tier requires reading failure/change/review history before use.

http | ts_port pandas-market-calendars-5.4.0-ts-port-2026-07-05 LOW ROT 8 CHECKSRECHECK 2026-08-18
calendar.exchange-sessionsmarketdata.alignmenttime-series.scheduling
POST /v1/tools/pandas-market-calendars/run with {calendar:'NYSE'|'NASDAQ'|'CME'|'LSE'|'XKRX',start_date:'YYYY-MM-DD',end_date:'YYYY-MM-DD'}; returns data_as_of, trading_days, sessions, holidays, early_closes, and conventions. Calendar data is generated from pandas_market_calendars 5.4.0 and is not live.
CHANGES /v1/cards/card_pandas_market_calendars/changesREVOCATIONS /v1/revocations?card_id=card_pandas_market_calendars
OPEN CARD

pandas-ta

Technical analysis indicator library for pandas DataFrames, useful for feature generation and quick market-data transformations.
SANDBOXED

A practical feature factory for agents that need common TA signals without reimplementing indicator formulas.

APEX CARD V2

Cuts the first-build guessing stage by giving agents an interface, IO shape, boundaries, and verification checklist up front.

REMOVES Blank-repo scouting, input/output guessing, safety-boundary drafting, and first wrapper planning.RISK DATA-ONLY
Curated as a reusable Apex contract.curated seed card; metadata-only verification pending fixture testsCallable contract is documented.
NEXT ACTION REFERENCEREAD STEPS 9
TIER SANDBOXEDLICENSE SEE-UPSTREAMVERIFIED 2026-07-19RECEIPTS 0REVIEWS 0RUNNABLE NO
TOKENS EST. 135MARKDOWN /v1/cards/card_pandas_ta_indicators.mdORIGIN curated seed card; metadata-only verification pending fixture tests

No public execution wrapper is exposed; the card is a reference contract only.

python_import unverified-latest MED ROT 4 CHECKSRECHECK 2026-08-18
features.technical-indicatorsmarketdata.transformresearch.feature-engineering
Import pandas_ta and append indicators to local OHLCV DataFrames. No trading action should be inferred from indicator output alone.
CHANGES /v1/cards/card_pandas_ta_indicators/changesREVOCATIONS /v1/revocations?card_id=card_pandas_ta_indicators
OPEN CARD