smith6jt-cop
- 171 skills
- 0 followers
- 3 repo stars
- 2 weeks ago last updated
- ▌
- ▌
- ▌
- ▌
- ▌
- ▌
- ▌ Python Performance Patterns · smith6jt-copLow-risk Python performance optimization patterns with verified speedups
- ▌ Notebook Import Conventions · smith6jt-copCorrect import patterns and file paths for maxfuse Jupyter notebooks
- ▌ Agent Validation Experiment · smith6jt-copA/B testing infrastructure for validating Claude agent integration in RL training. Trigger when: (1) planning agent validation, (2) analyzing agent effectiveness, (3) deciding whether to enable agents, (4) understanding agent cost-benefit, (5) reviewing agent prompt design, (6) debugging agent response parsing, (7) notebook API key loading issues, (8) agent guardrails and phase gates.
- ▌ Backtest Capital Accounting · smith6jt-copCritical fix for backtest capital accounting when equity curves are inflated or drawdown metrics are meaningless
- ▌ Branch Integration Workflow · smith6jt-copSafely review and merge remote branches into stable. Trigger when: (1) reviewing open PRs, (2) integrating feature branches, (3) checking for stale branches with potential conflicts.
- ▌ Capital Proportional Limits · smith6jt-copUse percentage-based capital limits, not hard-coded dollar amounts. Trigger when: (1) trades blocked on small accounts, (2) min_cash_buffer is absolute dollars, (3) capital allocation doesn't scale with account size, (4) fixed dollar thresholds in risk code.
- ▌ Dashboard Pnl Visualization · smith6jt-copTrading dashboard P&L visualization with profit tracker integration, win-rate overlays, R-multiples, and configurable settings
- ▌ Drawdown Guardrails Pattern · smith6jt-copConsistent drawdown control pattern for trading systems - backtests, live trading, and training
- ▌ Network Architecture Sizing · smith6jt-copPPO network architecture sizing for trading models. Trigger: (1) model files are unexpectedly small/large, (2) choosing hidden_dims for training, (3) balancing model capacity vs inference speed.
- ▌ Parallel Universe Selection · smith6jt-copParallel processing for universe selection with 3-4x speedup. Trigger when: (1) selection takes too long, (2) 'No data returned' warnings for cached symbols, (3) need to process 100+ candidates.
- ▌ Dashboard Feature Discovery · smith6jt-copSurface a shipped-but-undocumented CLI feature in user-facing docs. Trigger: user reports a known feature missing from README/readthedocs even though the CLI command exists.
- ▌ Notebook Module Refactoring · smith6jt-copSafe refactoring of Jupyter notebook code into Python modules
- ▌ Slurm Concurrent Processing · smith6jt-copKINTSUGI SLURM batch processing: Maximize throughput using multi-account resource calculation with GPU+CPU pools per account. Trigger: SLURM job submission, batch processing, resource maximization, GPU+CPU concurrent, headless processing, resource pool.
- ▌ Skills Registry Organization · smith6jt-copPattern for organizing Skills Registry as a shared submodule across projects
- ▌ Agent Validation Integration · smith6jt-copv4.1.0 integration of agent validation learnings into live trading, backtesting, and feedback loops
- ▌ Crypto Pair Fee Optimization · smith6jt-copAnalysis of crypto-to-crypto vs crypto-to-fiat (USD) trading on Alpaca for fee optimization and risk considerations
- ▌ Discounted Thompson Sampling · smith6jt-copUse when implementing Thompson sampling for non-stationary environments, handling regime changes in bandits, or tuning exploration-exploitation with decay
- ▌ Live Trader Prediction Fixes · smith6jt-copFix live trader prediction issues: zeros returned, excessive data fetching, insufficient bars. Trigger when predictions show all zeros, excessive API calls, or Markov features show uniform priors.
- ▌ Persistent Cache Gap Filling · smith6jt-copPersistent data cache with gap-filling for historical market data. Trigger when: (1) cache re-downloads complete data unnecessarily, (2) time-based cache expiry wastes API calls, (3) historical data needs incremental updates only, (4) gap-fill warnings appear when cache is sufficient.
- ▌ Position Sizing Action Space · smith6jt-cop7-action space with position sizing (25/50/75%) + small account simulation. Trigger when: (1) model needs sizing decisions, (2) training for <$25K accounts, (3) upgrading obs_dim 5600->5900.
- ▌ Prediction Outcome Recording · smith6jt-copRecord prediction outcomes when positions close. Trigger when: (1) predictions logged but outcomes missing, (2) calculating prediction accuracy, (3) tracking model performance.
- ▌ Private Function Shim Export · smith6jt-copFix compatibility shims that fail to export private functions. Trigger when: (1) import errors for _prefixed functions after module reorganization, (2) 'cannot import name' errors from shim files, (3) creating backwards-compat shims for moved modules.
- ▌ Selection Progress Reporting · smith6jt-copProgress reporting for universe selection phases. Trigger when: (1) selection appears to stall/hang, (2) no output after data fetching, (3) need visibility into hard filter or scoring phases.
- ▌ Symbol Selection Statistical · smith6jt-copStatistical analysis for selecting trading symbols compatible with mean-reversion/momentum strategies, GARCH risk models, and Markov regime detection
- ▌ Trading Gates Pattern Filter · smith6jt-copPattern filter normalization and multi-gate order entry system
- ▌ Batch Staging Rsync Patterns · smith6jt-copRsync patterns for staging CODEX raw data from orange to blue storage, avoiding cycle directory flattening and bash subshell variable loss
- ▌ Claude Md Context Management · smith6jt-copSplit oversized CLAUDE.md into subdirectory files. Trigger: CLAUDE.md too large, context window, documentation splitting.
- ▌ Lightsheet Psf Deconvolution · smith6jt-copKINTSUGI deconvolution: True lightsheet PSF calculation to fix horizontal banding artifacts. Trigger: deconvolution artifacts, horizontal banding, PSF issues, fcyl, slitwidth, lightsheet, LsDeconv.
- ▌ Tile Snake Pattern Stitching · smith6jt-copSNAKE vs RASTER tile acquisition patterns for image stitching in KINTSUGI
- ▌ Notebook Checkpoint Pattern · smith6jt-copPattern for connecting Jupyter notebooks in a pipeline via checkpoint files. Trigger: connecting notebooks, saving outputs, multi-notebook pipelines
- ▌ Backtest Observation Mismatch · smith6jt-copBacktest observation features that receive hardcoded defaults vs training dynamic values — and why fixing them DEGRADES results. Includes walk-forward batch mode fix.
- ▌ Openbb Enrichment Integration · smith6jt-copOpenBB Platform as parallel enrichment layer — fundamentals, FRED, earnings, news, options IV
- ▌ Codex Experiment JSON Parsing · smith6jt-copCODEX/Akoya experiment.json uses different field names and flat wavelength lists that must be translated to KINTSUGI format
- ▌ Genotype Hypothesis Analysis · smith6jt-copMulti-hypothesis genotype analysis framework for spatial biology data with small sample sizes. Trigger: (1) comparing measurements across genotype groups with small n (3-4/group), (2) multi-notebook analysis of spatial/multiplex imaging annotations, (3) vessel/region density normalized across tissue compartments.
- ▌ Conda Multi Account Hipergator · smith6jt-copSet up conda with multiple group storage locations on HiPerGator, including cache isolation
- ▌ Codebase Consolidation Pattern · smith6jt-copReorganize scattered modules into subpackages while maintaining backwards compatibility via shims. Trigger when: (1) too many root-level files, (2) need to group related functionality, (3) cleaning up codebase structure.
- ▌ Per Sector Candidate Filtering · smith6jt-copPer-sector candidate limits using yfinance sector data. Trigger when: (1) universe selection has too many candidates, (2) need sector diversity in candidate pool, (3) downloading historical data takes too long.
- ▌ Selection Diversity Validation · smith6jt-copFAIL LOUDLY pattern for selection constraints. Trigger when: (1) correlated stocks selected together, (2) min_crypto_positions ignored, (3) constraints violated silently.
- ▌ Symbol Selection Asset Filters · smith6jt-copFix symbol selection failing due to asset-type filter mismatch. Trigger when: (1) '0/N passed hard filter' error, (2) crypto selection fails unexpectedly, (3) assets filtered by wrong price/volume thresholds.
- ▌ Weighted Autofluorescence Subtraction · smith6jt-copMulti-range weighted autofluorescence subtraction for multiplex IF. Protects dim signal while aggressively removing bright AF. Trigger: autofluorescence removal, blank subtraction, dim marker preservation, weighted subtraction.
- ▌ Backtest Datetime Visualization · smith6jt-copConverting backtest visualizations from bar indices/timesteps to actual datetime axes for clearer time context
- ▌ Notebook Config Drift Detection · smith6jt-copDetect and fix config drift between Colab notebooks and GPUEnvConfig defaults. Triggers when: (1) reviewing notebooks before training, (2) updating vectorized_env.py defaults, (3) preparing A/B experiments, (4) debugging unexpected training behavior.
- ▌ Deconvolution Intensity Scaling · smith6jt-copKINTSUGI deconvolution: Output scaling must preserve original intensity relationships for quantitative analysis. Trigger: deconvolution saturates at 65535, channel intensity ratios inverted, hist_clip scaling issues.
- ▌ Snakemake Workflow Architecture · smith6jt-copSnakemake workflow design for KINTSUGI SLURM processing: lambda resource routing, GPU-only cycle pre-assignment, multi-account scheduling, registration + QC aggregate rules. Trigger: Snakemake rules, workflow config, ruleorder, cycle assignment, DAG design, sentinel files, workflow run/check, registration, VALIS, QC reports.
- ▌ Optional Dependency Test Mocking · smith6jt-copPattern for testing modules with optional dependencies by mocking at sys.modules level
- ▌ Plotly Figurewidget Interactive · smith6jt-copFix for Plotly FigureWidget KeyError: 'uid' when updating shapes with sliders. Trigger: interactive Plotly dashboard errors, FigureWidget batch_update issues
- ▌ Sparse Expression Visualization · smith6jt-copVisualization patterns for sparse single-cell gene expression data. Trigger: boxplots flat at zero, single-cell expression plots, sparse data visualization
- ▌ Crypto Hard Filter Simplification · smith6jt-copSimplify crypto hard filters to essential checks only. Trigger when: (1) crypto symbols fail multiple filters, (2) data_quality/spread/trading_status fail for crypto, (3) yfinance data gaps causing false failures.
- ▌ Testing Benchmarking Infrastructure · smith6jt-copTesting suite with benchmarking for version-to-version performance comparison
- ▌ Dashboard Symbol Discovery · smith6jt-copAuto-discover dashboard symbols from loaded RL models. Trigger when: (1) dashboard shows old/wrong symbols, (2) symbols mismatch between live trader and dashboard, (3) adding new models to system, (4) dashboard shows NO_MODEL for all symbols.
- ▌ Reward Scaling Calibration · smith6jt-copFix phantom MaxDD values in training by calibrating reward_scale. Trigger when: (1) validation MaxDD shows 35-80% values, (2) MaxDD doesn't correlate with training quality, (3) gating thresholds seem too lenient/strict.
- ▌ Training Code Verification · smith6jt-copChecklist for editing training code. Trigger when: (1) modifying training notebook, (2) changing GPUEnvConfig, (3) changing reward functions, (4) ANY training-related code changes.
- ▌ Training Improvements V245 · smith6jt-copTraining improvements: LR warmup, validation intervals, reward weights. Trigger when: (1) training unstable in early epochs, (2) need more validation visibility, (3) model too conservative.
- ▌ Trust Rl Model Predictions · smith6jt-copRemove pattern filters that second-guess RL model predictions. Trigger when: (1) signals show past threshold but status is HOLD, (2) pattern filter blocks model signals, (3) external gates contradict trained model, (4) dashboard shows disconnect between signal strength and action.
- ▌ Vix Scaled Risk Management · smith6jt-copVIX-based position sizing, circuit breakers, and capital shift suppression
- ▌ Live Aware Account Routing · smith6jt-copKINTSUGI Snakefile + CLI changes that route SLURM jobs around accounts saturated by OTHER users on the same QOS pool. Trigger: QOSGrpMemLimit, jobs stuck pending despite available GPU slots in config, noisy neighbor on shared QOS, multi-user investment pool exhaustion, _build_cycle_assignment static-vs-live.
- ▌ Project Init Data Handling · smith6jt-copKINTSUGI project initialization differentiates raw vs processed data and handles --slurm on existing projects
- ▌ Repo Project Sync Workflow · smith6jt-copWhen editing KINTSUGI notebook modules (Kdecon, Kstitch, Kreg, etc.), always edit the main repo first then sync to project folders
- ▌ Slurm Workflow Integration · smith6jt-copIntegrating SLURM job submission with kintsugi init for HPC clusters. Trigger: adding SLURM support to projects, auto-detecting HPC settings, CLI batch submission
- ▌ Snakemake Signal Isolation · smith6jt-copAdding signal isolation as a Snakemake pipeline rule for end-to-end automation. Trigger: signal isolation rule, pipeline integration, batch completion sentinel, autofluorescence subtraction as workflow step.
- ▌ Tissuumaps Export Pipeline · smith6jt-copTissUUmaps export pipeline: DZI tile conversion via pyvips, .tmap project file generation, rsync deploy, skip-existing manifest. Trigger: TissUUmaps, DZI, dzsave, tile pyramid, export, deploy, web visualization, OpenSeadragon, .tmap, marker color mapping.
- ▌ Unified Install Resolution · smith6jt-copFix cascading dependency conflicts from sequential pip install per group. Trigger: (1) `kintsugi install all` causes numpy/torch/CUDA version conflicts, (2) pip install per-group overwrites cross-group constraints, (3) dependency conflict cascade from sequential pip invocations
- ▌ Scvi Covariate Validation · smith6jt-copValidate scVI batch correction covariate choices by comparing latent space quality metrics
- ▌ Shiny Statistics Redesign · smith6jt-copStatistics tab redesign patterns: 5-section narrative layout, BH correction for stratified tests, plotly legend positioning, synthetic union row density fix, test type UX guidance
- ▌ Multiplex Imaging Analysis · smith6jt-copOptimizing segmentation feature extraction, SOM clustering, and spatial analysis for multiplex immunofluorescence imaging with cell compartment and ECM features
- ▌ Pytorch Common Pitfalls · smith6jt-copFixes common PyTorch bugs including percentile calculations, LayerNorm for Conv1d, and buffer edge cases in reinforcement learning and neural network code.
- ▌ Gpu Quality Priority · smith6jt-copKINTSUGI NOTEBOOK processing principles: Never sacrifice quality for speed, always use GPU when available. Trigger: performance optimization, CPU/GPU choice, fast mode, quality vs speed in NOTEBOOKS (not SLURM).
- ▌ Slurm Workflow Guide · smith6jt-copComplete 10-step SLURM workflow from raw data to processed results in KINTSUGI
- ▌ Type Checking Pattern · smith6jt-copFix F821 undefined name errors for optional dependencies using TYPE_CHECKING
- ▌ Region Aware Matching · smith6jt-copSpatial region-aware cell matching for CODEX/scRNAseq integration
- ▌ Agent Validation V420 · smith6jt-copAgent validation overhaul: reward weight overrides, fitness decline gate, pinned data, staged experiments
- ▌ Agent Validation V430 · smith6jt-copAgent validation v4.3.0 — Make agents act effectively by disabling harmful actions, lowering gates, and injecting cross-run learning
- ▌ Alpaca Algo Plus Data · smith6jt-copUse Alpaca Algo Trader Plus for 4+ years of historical data. Trigger when: (1) increasing lookback, (2) data source selection, (3) yfinance comparison.
- ▌ Notebook Config Audit · smith6jt-copAudit Jupyter notebooks for hardcoded values that contradict configuration cells. Trigger when: (1) notebook behavior differs from documented settings, (2) updating notebook version, (3) finding inconsistent values across cells.
- ▌
- ▌ Shiny Modularization · smith6jt-copPatterns and pitfalls for modularizing large Shiny apps with plotly, sf, and cross-module reactives
- ▌ Dataclass Optimization · smith6jt-copPython dataclass best practices: slots, frozen, validation. Trigger when optimizing dataclasses or creating config classes.
- ▌ Dependency Deprecation · smith6jt-copPattern for deprecating external dependencies while maintaining backward compatibility. Trigger: removing Java/FIJI/MATLAB, deprecating optional deps
- ▌ Account Aware Training · smith6jt-copAdd account state (P&L, win rate, drawdown) to RL observations + drawdown penalty in rewards. Trigger when: (1) model needs account awareness, (2) training should penalize drawdowns, (3) upgrading obs_dim 5300→5600.
- ▌ Adaptive Predator Prey · smith6jt-copUse when tuning predator-prey dynamics, regime detection coefficients, or cooldown mechanisms
- ▌ Extended Hours Trading · smith6jt-copExtended hours trading support for pre-market and after-hours sessions via Alpaca
- ▌ Markov Regime Features · smith6jt-copDebugging constant Markov regime features in RL observations - when HMM probabilities show uniform values instead of dynamic regime estimates
- ▌ Model Version Protocol · smith6jt-copModel-trader version compatibility protocol: Embed version metadata in checkpoints, validate at load time. Trigger when: (1) training and live trading versions diverge, (2) models fail to load, (3) action interpretation issues.
- ▌ Post Training Workflow · smith6jt-copPost-training model validation workflow: gating, backtesting, walk-forward validation, deployment decisions. Trigger after GPU training completes.
- ▌ Selection Data Caching · smith6jt-copSUPERSEDED by persistent-cache-gap-filling (v2.8.0). Cache data during symbol selection for instant repeat runs.
- ▌ Globus Dataset Staging · smith6jt-copGlobus CLI workflow for staging HuBMAP CODEX datasets from remote endpoints to HiPerGator
- ▌ Hdl73 Signal Isolation · smith6jt-copBatch signal isolation for Phenocycler spleen channels using matched blank pairs and kintsugi.signal
- ▌ Vessel3d Frangi Bugfix · smith6jt-copFrangi vesselness 3D filter: Ra formula correction, scikit-image deprecations, GPU VRAM guard
- ▌
- ▌ Bimodal Score Diagnosis · smith6jt-copDiagnosing and fixing bimodal matching score distributions in MaxFuse
- ▌ Agent Validation Review · smith6jt-copEnd-to-end review of agent validation system for live trader readiness. Audits gating thresholds, model health wiring, live feedback loop, notebook consistency.
- ▌
- ▌ Dotted Symbol Exclusion · smith6jt-copExclude ALL dotted symbols from yfinance sector lookups and training. Trigger when: (1) yfinance errors on warrants/units/class shares, (2) training notebook fails on excluded symbol types, (3) adding new symbol exclusion patterns.
- ▌ Gpu Correlation Caching · smith6jt-copGPU-accelerated correlation matrix computation with persistent SQLite caching to eliminate bottleneck at correlation calculation during symbol selection