Skills_Registry

Normalize long-form CODEX cycle folders to short form before notebooks run. Trigger: cyc001_reg001_*, hard-coded cyc paths breaking, staged CODEX raw data failing in Notebooks 1/2.

v5.6.0 joint multi-TF model: single model per symbol with broadcast 1Hour context replaces dual 15Min/1Hour models. Trigger: (1) replacing weighted-voting model aggregation, (2) adding broadcast features to vectorized env, (3) limited training data + worried about overfitting from doubling obs_dim, (4) backtest builder mismatch with newer feature counts.

DEPRECATED in v5.6.0 — see joint-multi-tf-v560 skill. Documents the v5.2.0 dual-model approach (train separate 15Min/1Hour models, combine via weighted voting). Still relevant for: (1) loading legacy v5.5.0 dual models, (2) understanding the historical aggregation layer, (3) resampling pattern via origin='start'.

Surface 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.

KINTSUGI 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.

KINTSUGI 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.

Use Alpaca CryptoDataStream websocket for real-time crypto bars in live trading (OHLC only, volume excluded)

Migrate improperly structured plugins to correct .claude-plugin/ layout

Pattern for testing modules with optional dependencies by mocking at sys.modules level

Fix 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

v4.1.0 integration of agent validation learnings into live trading, backtesting, and feedback loops

End-to-end review of agent validation system for live trader readiness. Audits gating thresholds, model health wiring, live feedback loop, notebook consistency.

v4.1.0 reward function redesign to fix overtrading and DSR dominance

Multi-timeframe backtesting combining 15Min + 1Hour model signals. Trigger when: (1) multi-TF backtest, (2) combining timeframe signals in backtest, (3) validating multi-TF strategy, (4) --multi-tf CLI flag.

Backtest observation features that receive hardcoded defaults vs training dynamic values — and why fixing them DEGRADES results. Includes walk-forward batch mode fix.

Statistics tab redesign patterns: 5-section narrative layout, BH correction for stratified tests, plotly legend positioning, synthetic union row density fix, test type UX guidance

Detect 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.

Persistent 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.

Patterns for adding interactive features to modular Shiny apps with H5AD data, conditional UI, embedded panels, tissue scatter, Leiden clustering

Patterns for H5AD-backed Shiny apps with Excel fallback, groovy data in .uns, scVI QC validation, Leiden clustering merge, per-donor tissue extraction

Validate scVI batch correction covariate choices by comparing latent space quality metrics

Skilled agent architecture replacing multi-agent system for RL training. Trigger when: (1) planning agent-guided training, (2) implementing tool-augmented LLM consultations, (3) comparing skilled vs multi-agent approaches, (4) designing simulate-verify loops for training, (5) implementing prompt evolution / learnable parameters, (6) understanding Claude Agent SDK integration in training, (7) debugging SkilledTrainer consultations or tool calls, (8) configuring agent safety bounds for training actions.

Bulk-update stale workflow scripts across KINTSUGI projects. Trigger: scripts not propagating, QC failures, batch stalls after code updates.

Multi-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.

Adding 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.

Batch signal isolation for Phenocycler spleen channels using matched blank pairs and kintsugi.signal

Convert single-channel TIF directories to pyramidal OME-TIFF with SubIFDs

Snakemake 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.

Fix MCP server 'Executable not found' by resolving absolute paths in .mcp.json

CPI/GDP/retail sales date awareness and VIX regime scaling for calendar features

OpenBB Platform as parallel enrichment layer — fundamentals, FRED, earnings, news, options IV

VIX-based position sizing, circuit breakers, and capital shift suppression

Detect processed data produced with old/buggy KINTSUGI code via timestamp forensics

Multi-project signal isolation with cascading recipe resolution

Agent validation v4.3.0 — Make agents act effectively by disabling harmful actions, lowering gates, and injecting cross-run learning

Frangi vesselness 3D filter: Ra formula correction, scikit-image deprecations, GPU VRAM guard

CCXT multi-exchange crypto broker integration for better liquidity during after-hours capital shift

Agent validation overhaul: reward weight overrides, fitness decline gate, pinned data, staged experiments

Split oversized CLAUDE.md into subdirectory files. Trigger: CLAUDE.md too large, context window, documentation splitting.

Post-training model validation workflow: gating, backtesting, walk-forward validation, deployment decisions. Trigger after GPU training completes.