mikeparcewski/ml
skills/data/ml/SKILL.md
This skill should be used when working with machine learning models — architecture review, training pipeline design, feature engineering, and deployment guidance. Use when: - "review this ML model" - "design ML training pipeline" - "how should I deploy this model" - "feature engineering advice" - "ML architecture guidance"
$ install --global
skillsauthnpx skillsauth add mikeparcewski/wicked-garden mlInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: May 7, 2026, 6:23 AM127.3s2 files scanned
SKILL.md
- name:
- ml
- description:
- |
- phase_relevance:
- ["design", "build"]
- archetype_relevance:
- ["*"]
ML Engineering Skill
Guide machine learning model development, training, and deployment.
Quick Start
Review Model Architecture
/wicked-garden:data:ml review path/to/model/
Reviews: Model choice, training data quality, evaluation strategy, deployment readiness.
Design Training Pipeline
/wicked-garden:data:ml pipeline --type classification
Generates: Data loading, feature engineering, training config, evaluation framework.
ML Development Workflow
1. Problem Definition
- [ ] Problem type clear (classification/regression/ranking)
- [ ] Success metric defined
- [ ] Baseline established
- [ ] Data availability confirmed
- [ ] Inference requirements understood (latency/throughput)
2. Data Assessment
- [ ] Sufficient volume (>1000x features)
- [ ] Labels accurate and consistent
- [ ] Features available at inference time
- [ ] No data leakage from target
- [ ] Class balance acceptable
3. Feature Engineering
Good features are: Predictive, Available at inference, Clean (no leakage), Interpretable.
Common transformations:
- Numeric: Scaling, log transform
- Categorical: One-hot, target encoding
- Time: Extract components, cyclical encoding
- Aggregations: Rolling windows, user stats
4. Model Selection
| Data Size | Structured | Recommendation | |-----------|------------|----------------| | <10K rows | Yes | Linear/Simple tree | | 10K-1M | Yes | GradientBoosting (XGBoost/LightGBM) | | >1M | Yes | Deep learning possible | | Any | Images/Text | Deep learning |
5. Training & Evaluation
Split strategy: Random (if i.i.d.), Time-based (if time series), Cross-validation (robust).
Key metrics:
- Classification: Accuracy, Precision, Recall, F1, AUC
- Regression: RMSE, MAE, R²
6. Hyperparameter Tuning
- Grid search: Exhaustive, slow
- Random search: More efficient
- Bayesian optimization: Most efficient
7. Deployment
Patterns: Batch scoring, REST API, Streaming
Checklist:
- [ ] Inference latency acceptable
- [ ] Model size appropriate
- [ ] Feature pipeline reproducible
- [ ] Monitoring instrumented
- [ ] Rollback plan documented
ML Monitoring
Model Performance: Prediction accuracy, distribution shifts, error rate by segment.
Data Quality: Feature distributions, missing rates, cardinality changes.
System Health: Latency (p50, p95, p99), throughput, memory.
Integration
- wicked-brain:search: Find model code with
wicked-brain:search "model|classifier"(FTS5 over indexed code) - Native tasks: Track experiments via TaskCreate with
metadata.event_type="task" - wicked-garden:data:analyze: Analyze training data via DuckDB SQL
Best Practices
- Always baseline: Start simple, measure improvement, justify complexity
- Avoid leakage: Use only past data, split before processing
- Monitor production: Track predictions, detect drift, plan retraining
- Document everything: Architecture, features, training data, results
Reference
For detailed techniques:
- Output Templates - Model review report, model card
Related Skills
mikeparcewski/wicked-garden-engineering-conformance-reviewer
development
VerifiedTrustedCommunity
Pattern-conformance agent-half: evaluates a produced artifact or diff against a set of architectural/design pattern rules from the conformance-rule store (wicked_governance schema). Returns structured findings with rule ID, severity, and rationale — the deterministic half (mechanical rule recall) is done by the guard pipeline; this is the semantic evaluation step. Triggered by: the guard_pipeline `outgov_pattern` check (session-close), or explicitly by an engineering review when WICKED_OUTGOV_RULES_DIR is populated. NOT a replacement for the full `engineering` review skill — focuses only on conformance to stored Pattern rules; architecture and code-quality checks live in the `engineering` skill. Semantic evaluation reuses `wicked-garden-qe-semantic-reviewer` as the designated agent-half evaluator (per garden#983 spec). This skill is the orchestrating wrapper that loads applicable Pattern rules and delegates the per-rule semantic judgment to qe-semantic-reviewer.
8SKILL.mdUpdated Jul 22, 2026
mikeparcewski/wicked-garden-domain
tools
VerifiedTrustedCommunity
The FOUNDATIONAL domain-model capability: extract a codebase's domain — testable business rules (with confidence + provenance), entities, requirements — as a schema-conformant model on the estate graph. The workers annotate the store; wicked-core reads it and builds the requirements graph, coverage-gating fail-closed. Steers three fork workers. A shared substrate, not a modernization tool. The `modernize` archetype DERIVES from it; build / migrate / review / specify / explore consume the SAME domain model — none OWN it. Understanding a codebase's domain is upstream of almost everything else garden does. Use when: "extract the business rules / domain model from this codebase", "build a requirements graph from the code", "what does this system actually require", "reverse-engineer the domain before we build/port/migrate". Works on ANY codebase (modern or legacy) — the value is the domain model, not the porting. NOT the code transform itself (that is the archetype consuming this model). This skill produces the DOMAIN MODEL, not new code.
8SKILL.mdUpdated Jul 15, 2026
mikeparcewski/wicked-garden-domain-modeler
development
VerifiedTrustedCommunity
Domain-graph fork worker for the modernize archetype. Groups the estate's Louvain communities into business domains, attaches each requirement to its cluster (advisory cluster_id provenance), and invokes wicked-core's domain-graph build (which reads the annotated estate store, recomputes coverage fail-closed, and builds the requirements graph) — then validates core's output against the vendored schema. Use when: dispatched by wicked-garden-domain after rule extraction to turn a flat rule set into cluster-keyed domains; "group these into domains", "build the requirements graph", "translate clusters into a domain model". NOT for mining the rules themselves (that is domain-extractor) or threat-modeling (that is domain-coverage).
8SKILL.mdUpdated Jul 15, 2026
mikeparcewski/wicked-garden-domain-extractor
tools
VerifiedTrustedCommunity
Rule-extraction fork worker for the FOUNDATIONAL domain-model capability. Mines testable business rules from a codebase — each with a numeric confidence and a provenance{source, ref, source_kinds} — and annotates them into the estate store so wicked-core can build the domain-model requirements graph (coverage-gated). This is a substrate, not a modernization tool: the `modernize` archetype DERIVES from it, and build / migrate / review / specify / explore can consume the same domain model — none OWN it. Use when: dispatched by wicked-garden-domain to mine the business_rules of a codebase (or a module); "extract the domain rules", "what does this system require", building the requirements half of a domain model. NOT for grouping into domains (that is domain-modeler) or judging coverage (that is domain-coverage — a seat-distinct evaluator).
8SKILL.mdUpdated Jul 15, 2026