[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-databricks-databricks-mlflow-evaluation":3,"mdc--ovvhc7-key":36,"related-repo-databricks-databricks-mlflow-evaluation":2037,"related-org-databricks-databricks-mlflow-evaluation":2154},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":26,"repoUrl":27,"updatedAt":28,"license":29,"forks":30,"topics":31,"repo":32,"sourceUrl":34,"mdContent":35},"databricks-mlflow-evaluation","evaluate GenAI agents with MLflow","MLflow 3 GenAI agent evaluation. Use when writing mlflow.genai.evaluate() code, creating @scorer functions, using built-in scorers (Guidelines, Correctness, Safety, RetrievalGroundedness), building eval datasets from traces, setting up trace ingestion and production monitoring, aligning judges with MemAlign from domain expert feedback, or running optimize_prompts() with GEPA for automated prompt improvement.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},"databricks","Databricks","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fdatabricks.png",[12,16,19,22,23],{"name":13,"slug":14,"type":15},"LLM","llm","tag",{"name":17,"slug":18,"type":15},"MLflow","mlflow",{"name":20,"slug":21,"type":15},"Evals","evals",{"name":9,"slug":8,"type":15},{"name":24,"slug":25,"type":15},"Agents","agents",204,"https:\u002F\u002Fgithub.com\u002Fdatabricks\u002Fdatabricks-agent-skills","2026-07-12T08:04:07.451169",null,60,[],{"repoUrl":27,"stars":26,"forks":30,"topics":33,"description":29},[],"https:\u002F\u002Fgithub.com\u002Fdatabricks\u002Fdatabricks-agent-skills\u002Ftree\u002FHEAD\u002Fplugins\u002Fdatabricks\u002Fcopilot\u002Fskills\u002Fdatabricks-mlflow-evaluation","---\nname: databricks-mlflow-evaluation\ndescription: \"MLflow 3 GenAI agent evaluation. Use when writing mlflow.genai.evaluate() code, creating @scorer functions, using built-in scorers (Guidelines, Correctness, Safety, RetrievalGroundedness), building eval datasets from traces, setting up trace ingestion and production monitoring, aligning judges with MemAlign from domain expert feedback, or running optimize_prompts() with GEPA for automated prompt improvement.\"\ncompatibility: Requires databricks CLI (>= v1.0.0)\nmetadata:\n  version: \"0.1.0\"\nparent: databricks-core\n---\n\n# MLflow 3 GenAI Evaluation\n\n## Scope vs upstream `mlflow\u002Fskills`\n\nThe OSS `mlflow\u002Fskills` repo ships [`agent-evaluation`](https:\u002F\u002Fgithub.com\u002Fmlflow\u002Fskills\u002Ftree\u002Fmain\u002Fagent-evaluation) and related skills (`instrumenting-with-mlflow-tracing`, `analyze-mlflow-trace`, `retrieving-mlflow-traces`, `querying-mlflow-metrics`) that cover the generic MLflow GenAI evaluation workflow — `mlflow.genai.evaluate()`, scorers\u002Fjudges, datasets, tracing setup, and the 5-step evaluation loop.\n\nThis skill **layers Databricks-specific patterns on top of that workflow** rather than restating it. Use this skill when you need any of:\n\n- **Unity Catalog trace ingestion** — production traces written into UC tables, log-based monitoring (`patterns-trace-ingestion.md`).\n- **MemAlign judge alignment via UC SME labeling sessions** — aligning custom judges against domain-expert feedback collected in Databricks (`patterns-judge-alignment.md`).\n- **`optimize_prompts()` GEPA loop** — Databricks' automated prompt-optimization driver running on a UC dataset (`patterns-prompt-optimization.md`).\n- **Databricks-flavored scorer\u002Fdataset patterns** — UC-table-backed datasets, tagging traces in the Databricks UI for inclusion (`patterns-datasets.md`, `patterns-scorers.md`).\n\nFor everything else — generic `mlflow.genai.evaluate()` calls, scorer authoring patterns, dataset creation outside Databricks, MLflow tracing setup that isn't UC-table-bound — the upstream `mlflow\u002Fskills\u002Fagent-evaluation` skill is the canonical source and is kept current by the MLflow team.\n\n## Before Writing Any Code\n\n1. **Read GOTCHAS.md** - 15+ common mistakes that cause failures\n2. **Read CRITICAL-interfaces.md** - Exact API signatures and data schemas\n\n## End-to-End Workflows\n\nFollow these workflows based on your goal. Each step indicates which reference files to read.\n\n### Workflow 1: First-Time Evaluation Setup\n\nFor users new to MLflow GenAI evaluation or setting up evaluation for a new agent.\n\n| Step | Action | Reference Files |\n|------|--------|-----------------|\n| 1 | Understand what to evaluate | `user-journeys.md` (Journey 0: Strategy) |\n| 2 | Learn API patterns | `GOTCHAS.md` + `CRITICAL-interfaces.md` |\n| 3 | Build initial dataset | `patterns-datasets.md` (Patterns 1-4) |\n| 4 | Choose\u002Fcreate scorers | `patterns-scorers.md` + `CRITICAL-interfaces.md` (built-in list) |\n| 5 | Run evaluation | `patterns-evaluation.md` (Patterns 1-3) |\n\n### Workflow 2: Production Trace -> Evaluation Dataset\n\nFor building evaluation datasets from production traces.\n\n| Step | Action | Reference Files |\n|------|--------|-----------------|\n| 1 | Search and filter traces | `patterns-trace-analysis.md` (MCP tools section) |\n| 2 | Analyze trace quality | `patterns-trace-analysis.md` (Patterns 1-7) |\n| 3 | Tag traces for inclusion | `patterns-datasets.md` (Patterns 16-17) |\n| 4 | Build dataset from traces | `patterns-datasets.md` (Patterns 6-7) |\n| 5 | Add expectations\u002Fground truth | `patterns-datasets.md` (Pattern 2) |\n\n### Workflow 3: Performance Optimization\n\nFor debugging slow or expensive agent execution.\n\n| Step | Action | Reference Files |\n|------|--------|-----------------|\n| 1 | Profile latency by span | `patterns-trace-analysis.md` (Patterns 4-6) |\n| 2 | Analyze token usage | `patterns-trace-analysis.md` (Pattern 9) |\n| 3 | Detect context issues | `patterns-context-optimization.md` (Section 5) |\n| 4 | Apply optimizations | `patterns-context-optimization.md` (Sections 1-4, 6) |\n| 5 | Re-evaluate to measure impact | `patterns-evaluation.md` (Pattern 6-7) |\n\n### Workflow 4: Regression Detection\n\nFor comparing agent versions and finding regressions.\n\n| Step | Action | Reference Files |\n|------|--------|-----------------|\n| 1 | Establish baseline | `patterns-evaluation.md` (Pattern 4: named runs) |\n| 2 | Run current version | `patterns-evaluation.md` (Pattern 1) |\n| 3 | Compare metrics | `patterns-evaluation.md` (Patterns 6-7) |\n| 4 | Analyze failing traces | `patterns-trace-analysis.md` (Pattern 7) |\n| 5 | Debug specific failures | `patterns-trace-analysis.md` (Patterns 8-9) |\n\n### Workflow 5: Custom Scorer Development\n\nFor creating project-specific evaluation metrics.\n\n| Step | Action | Reference Files |\n|------|--------|-----------------|\n| 1 | Understand scorer interface | `CRITICAL-interfaces.md` (Scorer section) |\n| 2 | Choose scorer pattern | `patterns-scorers.md` (Patterns 4-11) |\n| 3 | For multi-agent scorers | `patterns-scorers.md` (Patterns 13-16) |\n| 4 | Test with evaluation | `patterns-evaluation.md` (Pattern 1) |\n\n### Workflow 6: Unity Catalog Trace Ingestion & Production Monitoring\n\nFor storing traces in Unity Catalog, instrumenting applications, and enabling continuous production monitoring.\n\n| Step | Action | Reference Files |\n|------|--------|-----------------|\n| 1 | Link UC schema to experiment | `patterns-trace-ingestion.md` (Patterns 1-2) |\n| 2 | Set trace destination | `patterns-trace-ingestion.md` (Patterns 3-4) |\n| 3 | Instrument your application | `patterns-trace-ingestion.md` (Patterns 5-8) |\n| 4 | Configure trace sources (Apps\u002FServing\u002FOTEL) | `patterns-trace-ingestion.md` (Patterns 9-11) |\n| 5 | Enable production monitoring | `patterns-trace-ingestion.md` (Patterns 12-13) |\n| 6 | Query and analyze UC traces | `patterns-trace-ingestion.md` (Pattern 14) |\n\n### Workflow 7: Judge Alignment with MemAlign\n\nFor aligning an LLM judge to match domain expert preferences. A well-aligned judge improves every downstream use: evaluation accuracy, production monitoring signal, and prompt optimization quality. This workflow is valuable on its own, independent of prompt optimization.\n\n| Step | Action | Reference Files |\n|------|--------|-----------------|\n| 1 | Design base judge with `make_judge` (any feedback type) | `patterns-judge-alignment.md` (Pattern 1) |\n| 2 | Run evaluate(), tag successful traces | `patterns-judge-alignment.md` (Pattern 2) |\n| 3 | Build UC dataset + create SME labeling session | `patterns-judge-alignment.md` (Pattern 3) |\n| 4 | Align judge with MemAlign after labeling completes | `patterns-judge-alignment.md` (Pattern 4) |\n| 5 | Register aligned judge to experiment | `patterns-judge-alignment.md` (Pattern 5) |\n| 6 | Re-evaluate with aligned judge (baseline) | `patterns-judge-alignment.md` (Pattern 6) |\n\n### Workflow 8: Automated Prompt Optimization with GEPA\n\nFor automatically improving a registered system prompt using `optimize_prompts()`. Works with any scorer, but paired with an aligned judge (Workflow 7) gives the most domain-accurate signal. For the full end-to-end loop combining alignment and optimization, see `user-journeys.md` Journey 10.\n\n| Step | Action | Reference Files |\n|------|--------|-----------------|\n| 1 | Build optimization dataset (inputs + expectations) | `patterns-prompt-optimization.md` (Pattern 1) |\n| 2 | Run optimize_prompts() with GEPA + scorer | `patterns-prompt-optimization.md` (Pattern 2) |\n| 3 | Register new version, promote conditionally | `patterns-prompt-optimization.md` (Pattern 3) |\n\n## Reference Files Quick Lookup\n\n| Reference | Purpose | When to Read |\n|-----------|---------|--------------|\n| `GOTCHAS.md` | Common mistakes | **Always read first** before writing code |\n| `CRITICAL-interfaces.md` | API signatures, schemas | When writing any evaluation code |\n| `patterns-evaluation.md` | Running evals, comparing | When executing evaluations |\n| `patterns-scorers.md` | Custom scorer creation | When built-in scorers aren't enough |\n| `patterns-datasets.md` | Dataset building | When preparing evaluation data |\n| `patterns-trace-analysis.md` | Trace debugging | When analyzing agent behavior |\n| `patterns-context-optimization.md` | Token\u002Flatency fixes | When agent is slow or expensive |\n| `patterns-trace-ingestion.md` | UC trace setup, monitoring | When setting up trace storage or production monitoring |\n| `patterns-judge-alignment.md` | MemAlign judge alignment, labeling sessions, SME feedback | When aligning judges to domain expert preferences |\n| `patterns-prompt-optimization.md` | GEPA optimization: build dataset, optimize_prompts(), promote | When running automated prompt improvement |\n| `user-journeys.md` | High-level workflows, full domain-expert optimization loop | When starting a new evaluation project or running the full align + optimize cycle |\n\n## Critical API Facts\n\n- **Use:** `mlflow.genai.evaluate()` (NOT `mlflow.evaluate()`)\n- **Data format:** `{\"inputs\": {\"query\": \"...\"}}` (nested structure required)\n- **predict_fn:** Receives `**unpacked kwargs` (not a dict)\n- **MemAlign:** Scorer-agnostic (works with any `feedback_value_type` -- float, bool, categorical); token-heavy on the embedding model so set `embedding_model` explicitly\n- **Label schema name matching:** The label schema `name` in the labeling session MUST match the judge `name` used in `evaluate()` for `align()` to pair scores\n- **Aligned judge scores:** May be lower than unaligned judge scores -- this is expected and means the judge is now more accurate, not that the agent regressed\n- **GEPA optimization dataset:** Must have both `inputs` AND `expectations` per record (different from eval dataset)\n- **Episodic memory:** Lazily loaded -- `get_scorer()` results won't show episodic memory on print until the judge is first used\n- **optimize_prompts:** Requires MLflow >= 3.5.0\n\nSee `GOTCHAS.md` for complete list.\n\n## Related Skills\n\n- **[databricks-docs](..\u002Fdatabricks-docs\u002FSKILL.md)** - General Databricks documentation reference\n- **databricks-model-serving** - Deploying models and agents to serving endpoints\n- **[databricks-agent-bricks](..\u002Fdatabricks-agent-bricks\u002FSKILL.md)** - Building agents that can be evaluated with this skill\n- **[databricks-python-sdk](..\u002Fdatabricks-python-sdk\u002FSKILL.md)** - SDK patterns used alongside MLflow APIs\n- **[databricks-unity-catalog](..\u002Fdatabricks-unity-catalog\u002FSKILL.md)** - Unity Catalog tables for managed evaluation datasets\n",{"data":37,"body":42},{"name":4,"description":6,"compatibility":38,"metadata":39,"parent":41},"Requires databricks CLI (>= v1.0.0)",{"version":40},"0.1.0","databricks-core",{"type":43,"children":44},"root",[45,54,68,134,147,234,254,260,284,290,295,302,307,470,476,481,616,622,627,762,768,773,906,912,917,1028,1034,1039,1196,1202,1207,1369,1375,1394,1481,1487,1750,1756,1950,1962,1968],{"type":46,"tag":47,"props":48,"children":50},"element","h1",{"id":49},"mlflow-3-genai-evaluation",[51],{"type":52,"value":53},"text","MLflow 3 GenAI Evaluation",{"type":46,"tag":55,"props":56,"children":58},"h2",{"id":57},"scope-vs-upstream-mlflowskills",[59,61],{"type":52,"value":60},"Scope vs upstream ",{"type":46,"tag":62,"props":63,"children":65},"code",{"className":64},[],[66],{"type":52,"value":67},"mlflow\u002Fskills",{"type":46,"tag":69,"props":70,"children":71},"p",{},[72,74,79,81,94,96,102,104,110,111,117,118,124,126,132],{"type":52,"value":73},"The OSS ",{"type":46,"tag":62,"props":75,"children":77},{"className":76},[],[78],{"type":52,"value":67},{"type":52,"value":80}," repo ships ",{"type":46,"tag":82,"props":83,"children":87},"a",{"href":84,"rel":85},"https:\u002F\u002Fgithub.com\u002Fmlflow\u002Fskills\u002Ftree\u002Fmain\u002Fagent-evaluation",[86],"nofollow",[88],{"type":46,"tag":62,"props":89,"children":91},{"className":90},[],[92],{"type":52,"value":93},"agent-evaluation",{"type":52,"value":95}," and related skills (",{"type":46,"tag":62,"props":97,"children":99},{"className":98},[],[100],{"type":52,"value":101},"instrumenting-with-mlflow-tracing",{"type":52,"value":103},", ",{"type":46,"tag":62,"props":105,"children":107},{"className":106},[],[108],{"type":52,"value":109},"analyze-mlflow-trace",{"type":52,"value":103},{"type":46,"tag":62,"props":112,"children":114},{"className":113},[],[115],{"type":52,"value":116},"retrieving-mlflow-traces",{"type":52,"value":103},{"type":46,"tag":62,"props":119,"children":121},{"className":120},[],[122],{"type":52,"value":123},"querying-mlflow-metrics",{"type":52,"value":125},") that cover the generic MLflow GenAI evaluation workflow — ",{"type":46,"tag":62,"props":127,"children":129},{"className":128},[],[130],{"type":52,"value":131},"mlflow.genai.evaluate()",{"type":52,"value":133},", scorers\u002Fjudges, datasets, tracing setup, and the 5-step evaluation loop.",{"type":46,"tag":69,"props":135,"children":136},{},[137,139,145],{"type":52,"value":138},"This skill ",{"type":46,"tag":140,"props":141,"children":142},"strong",{},[143],{"type":52,"value":144},"layers Databricks-specific patterns on top of that workflow",{"type":52,"value":146}," rather than restating it. Use this skill when you need any of:",{"type":46,"tag":148,"props":149,"children":150},"ul",{},[151,170,187,210],{"type":46,"tag":152,"props":153,"children":154},"li",{},[155,160,162,168],{"type":46,"tag":140,"props":156,"children":157},{},[158],{"type":52,"value":159},"Unity Catalog trace ingestion",{"type":52,"value":161}," — production traces written into UC tables, log-based monitoring (",{"type":46,"tag":62,"props":163,"children":165},{"className":164},[],[166],{"type":52,"value":167},"patterns-trace-ingestion.md",{"type":52,"value":169},").",{"type":46,"tag":152,"props":171,"children":172},{},[173,178,180,186],{"type":46,"tag":140,"props":174,"children":175},{},[176],{"type":52,"value":177},"MemAlign judge alignment via UC SME labeling sessions",{"type":52,"value":179}," — aligning custom judges against domain-expert feedback collected in Databricks (",{"type":46,"tag":62,"props":181,"children":183},{"className":182},[],[184],{"type":52,"value":185},"patterns-judge-alignment.md",{"type":52,"value":169},{"type":46,"tag":152,"props":188,"children":189},{},[190,201,203,209],{"type":46,"tag":140,"props":191,"children":192},{},[193,199],{"type":46,"tag":62,"props":194,"children":196},{"className":195},[],[197],{"type":52,"value":198},"optimize_prompts()",{"type":52,"value":200}," GEPA loop",{"type":52,"value":202}," — Databricks' automated prompt-optimization driver running on a UC dataset (",{"type":46,"tag":62,"props":204,"children":206},{"className":205},[],[207],{"type":52,"value":208},"patterns-prompt-optimization.md",{"type":52,"value":169},{"type":46,"tag":152,"props":211,"children":212},{},[213,218,220,226,227,233],{"type":46,"tag":140,"props":214,"children":215},{},[216],{"type":52,"value":217},"Databricks-flavored scorer\u002Fdataset patterns",{"type":52,"value":219}," — UC-table-backed datasets, tagging traces in the Databricks UI for inclusion (",{"type":46,"tag":62,"props":221,"children":223},{"className":222},[],[224],{"type":52,"value":225},"patterns-datasets.md",{"type":52,"value":103},{"type":46,"tag":62,"props":228,"children":230},{"className":229},[],[231],{"type":52,"value":232},"patterns-scorers.md",{"type":52,"value":169},{"type":46,"tag":69,"props":235,"children":236},{},[237,239,244,246,252],{"type":52,"value":238},"For everything else — generic ",{"type":46,"tag":62,"props":240,"children":242},{"className":241},[],[243],{"type":52,"value":131},{"type":52,"value":245}," calls, scorer authoring patterns, dataset creation outside Databricks, MLflow tracing setup that isn't UC-table-bound — the upstream ",{"type":46,"tag":62,"props":247,"children":249},{"className":248},[],[250],{"type":52,"value":251},"mlflow\u002Fskills\u002Fagent-evaluation",{"type":52,"value":253}," skill is the canonical source and is kept current by the MLflow team.",{"type":46,"tag":55,"props":255,"children":257},{"id":256},"before-writing-any-code",[258],{"type":52,"value":259},"Before Writing Any Code",{"type":46,"tag":261,"props":262,"children":263},"ol",{},[264,274],{"type":46,"tag":152,"props":265,"children":266},{},[267,272],{"type":46,"tag":140,"props":268,"children":269},{},[270],{"type":52,"value":271},"Read GOTCHAS.md",{"type":52,"value":273}," - 15+ common mistakes that cause failures",{"type":46,"tag":152,"props":275,"children":276},{},[277,282],{"type":46,"tag":140,"props":278,"children":279},{},[280],{"type":52,"value":281},"Read CRITICAL-interfaces.md",{"type":52,"value":283}," - Exact API signatures and data schemas",{"type":46,"tag":55,"props":285,"children":287},{"id":286},"end-to-end-workflows",[288],{"type":52,"value":289},"End-to-End Workflows",{"type":46,"tag":69,"props":291,"children":292},{},[293],{"type":52,"value":294},"Follow these workflows based on your goal. Each step indicates which reference files to read.",{"type":46,"tag":296,"props":297,"children":299},"h3",{"id":298},"workflow-1-first-time-evaluation-setup",[300],{"type":52,"value":301},"Workflow 1: First-Time Evaluation Setup",{"type":46,"tag":69,"props":303,"children":304},{},[305],{"type":52,"value":306},"For users new to MLflow GenAI evaluation or setting up evaluation for a new agent.",{"type":46,"tag":308,"props":309,"children":310},"table",{},[311,335],{"type":46,"tag":312,"props":313,"children":314},"thead",{},[315],{"type":46,"tag":316,"props":317,"children":318},"tr",{},[319,325,330],{"type":46,"tag":320,"props":321,"children":322},"th",{},[323],{"type":52,"value":324},"Step",{"type":46,"tag":320,"props":326,"children":327},{},[328],{"type":52,"value":329},"Action",{"type":46,"tag":320,"props":331,"children":332},{},[333],{"type":52,"value":334},"Reference Files",{"type":46,"tag":336,"props":337,"children":338},"tbody",{},[339,364,394,417,446],{"type":46,"tag":316,"props":340,"children":341},{},[342,348,353],{"type":46,"tag":343,"props":344,"children":345},"td",{},[346],{"type":52,"value":347},"1",{"type":46,"tag":343,"props":349,"children":350},{},[351],{"type":52,"value":352},"Understand what to evaluate",{"type":46,"tag":343,"props":354,"children":355},{},[356,362],{"type":46,"tag":62,"props":357,"children":359},{"className":358},[],[360],{"type":52,"value":361},"user-journeys.md",{"type":52,"value":363}," (Journey 0: Strategy)",{"type":46,"tag":316,"props":365,"children":366},{},[367,372,377],{"type":46,"tag":343,"props":368,"children":369},{},[370],{"type":52,"value":371},"2",{"type":46,"tag":343,"props":373,"children":374},{},[375],{"type":52,"value":376},"Learn API patterns",{"type":46,"tag":343,"props":378,"children":379},{},[380,386,388],{"type":46,"tag":62,"props":381,"children":383},{"className":382},[],[384],{"type":52,"value":385},"GOTCHAS.md",{"type":52,"value":387}," + ",{"type":46,"tag":62,"props":389,"children":391},{"className":390},[],[392],{"type":52,"value":393},"CRITICAL-interfaces.md",{"type":46,"tag":316,"props":395,"children":396},{},[397,402,407],{"type":46,"tag":343,"props":398,"children":399},{},[400],{"type":52,"value":401},"3",{"type":46,"tag":343,"props":403,"children":404},{},[405],{"type":52,"value":406},"Build initial dataset",{"type":46,"tag":343,"props":408,"children":409},{},[410,415],{"type":46,"tag":62,"props":411,"children":413},{"className":412},[],[414],{"type":52,"value":225},{"type":52,"value":416}," (Patterns 1-4)",{"type":46,"tag":316,"props":418,"children":419},{},[420,425,430],{"type":46,"tag":343,"props":421,"children":422},{},[423],{"type":52,"value":424},"4",{"type":46,"tag":343,"props":426,"children":427},{},[428],{"type":52,"value":429},"Choose\u002Fcreate scorers",{"type":46,"tag":343,"props":431,"children":432},{},[433,438,439,444],{"type":46,"tag":62,"props":434,"children":436},{"className":435},[],[437],{"type":52,"value":232},{"type":52,"value":387},{"type":46,"tag":62,"props":440,"children":442},{"className":441},[],[443],{"type":52,"value":393},{"type":52,"value":445}," (built-in list)",{"type":46,"tag":316,"props":447,"children":448},{},[449,454,459],{"type":46,"tag":343,"props":450,"children":451},{},[452],{"type":52,"value":453},"5",{"type":46,"tag":343,"props":455,"children":456},{},[457],{"type":52,"value":458},"Run evaluation",{"type":46,"tag":343,"props":460,"children":461},{},[462,468],{"type":46,"tag":62,"props":463,"children":465},{"className":464},[],[466],{"type":52,"value":467},"patterns-evaluation.md",{"type":52,"value":469}," (Patterns 1-3)",{"type":46,"tag":296,"props":471,"children":473},{"id":472},"workflow-2-production-trace-evaluation-dataset",[474],{"type":52,"value":475},"Workflow 2: Production Trace -> Evaluation Dataset",{"type":46,"tag":69,"props":477,"children":478},{},[479],{"type":52,"value":480},"For building evaluation datasets from production traces.",{"type":46,"tag":308,"props":482,"children":483},{},[484,502],{"type":46,"tag":312,"props":485,"children":486},{},[487],{"type":46,"tag":316,"props":488,"children":489},{},[490,494,498],{"type":46,"tag":320,"props":491,"children":492},{},[493],{"type":52,"value":324},{"type":46,"tag":320,"props":495,"children":496},{},[497],{"type":52,"value":329},{"type":46,"tag":320,"props":499,"children":500},{},[501],{"type":52,"value":334},{"type":46,"tag":336,"props":503,"children":504},{},[505,528,550,572,594],{"type":46,"tag":316,"props":506,"children":507},{},[508,512,517],{"type":46,"tag":343,"props":509,"children":510},{},[511],{"type":52,"value":347},{"type":46,"tag":343,"props":513,"children":514},{},[515],{"type":52,"value":516},"Search and filter traces",{"type":46,"tag":343,"props":518,"children":519},{},[520,526],{"type":46,"tag":62,"props":521,"children":523},{"className":522},[],[524],{"type":52,"value":525},"patterns-trace-analysis.md",{"type":52,"value":527}," (MCP tools section)",{"type":46,"tag":316,"props":529,"children":530},{},[531,535,540],{"type":46,"tag":343,"props":532,"children":533},{},[534],{"type":52,"value":371},{"type":46,"tag":343,"props":536,"children":537},{},[538],{"type":52,"value":539},"Analyze trace quality",{"type":46,"tag":343,"props":541,"children":542},{},[543,548],{"type":46,"tag":62,"props":544,"children":546},{"className":545},[],[547],{"type":52,"value":525},{"type":52,"value":549}," (Patterns 1-7)",{"type":46,"tag":316,"props":551,"children":552},{},[553,557,562],{"type":46,"tag":343,"props":554,"children":555},{},[556],{"type":52,"value":401},{"type":46,"tag":343,"props":558,"children":559},{},[560],{"type":52,"value":561},"Tag traces for inclusion",{"type":46,"tag":343,"props":563,"children":564},{},[565,570],{"type":46,"tag":62,"props":566,"children":568},{"className":567},[],[569],{"type":52,"value":225},{"type":52,"value":571}," (Patterns 16-17)",{"type":46,"tag":316,"props":573,"children":574},{},[575,579,584],{"type":46,"tag":343,"props":576,"children":577},{},[578],{"type":52,"value":424},{"type":46,"tag":343,"props":580,"children":581},{},[582],{"type":52,"value":583},"Build dataset from traces",{"type":46,"tag":343,"props":585,"children":586},{},[587,592],{"type":46,"tag":62,"props":588,"children":590},{"className":589},[],[591],{"type":52,"value":225},{"type":52,"value":593}," (Patterns 6-7)",{"type":46,"tag":316,"props":595,"children":596},{},[597,601,606],{"type":46,"tag":343,"props":598,"children":599},{},[600],{"type":52,"value":453},{"type":46,"tag":343,"props":602,"children":603},{},[604],{"type":52,"value":605},"Add expectations\u002Fground truth",{"type":46,"tag":343,"props":607,"children":608},{},[609,614],{"type":46,"tag":62,"props":610,"children":612},{"className":611},[],[613],{"type":52,"value":225},{"type":52,"value":615}," (Pattern 2)",{"type":46,"tag":296,"props":617,"children":619},{"id":618},"workflow-3-performance-optimization",[620],{"type":52,"value":621},"Workflow 3: Performance Optimization",{"type":46,"tag":69,"props":623,"children":624},{},[625],{"type":52,"value":626},"For debugging slow or expensive agent execution.",{"type":46,"tag":308,"props":628,"children":629},{},[630,648],{"type":46,"tag":312,"props":631,"children":632},{},[633],{"type":46,"tag":316,"props":634,"children":635},{},[636,640,644],{"type":46,"tag":320,"props":637,"children":638},{},[639],{"type":52,"value":324},{"type":46,"tag":320,"props":641,"children":642},{},[643],{"type":52,"value":329},{"type":46,"tag":320,"props":645,"children":646},{},[647],{"type":52,"value":334},{"type":46,"tag":336,"props":649,"children":650},{},[651,673,695,718,740],{"type":46,"tag":316,"props":652,"children":653},{},[654,658,663],{"type":46,"tag":343,"props":655,"children":656},{},[657],{"type":52,"value":347},{"type":46,"tag":343,"props":659,"children":660},{},[661],{"type":52,"value":662},"Profile latency by span",{"type":46,"tag":343,"props":664,"children":665},{},[666,671],{"type":46,"tag":62,"props":667,"children":669},{"className":668},[],[670],{"type":52,"value":525},{"type":52,"value":672}," (Patterns 4-6)",{"type":46,"tag":316,"props":674,"children":675},{},[676,680,685],{"type":46,"tag":343,"props":677,"children":678},{},[679],{"type":52,"value":371},{"type":46,"tag":343,"props":681,"children":682},{},[683],{"type":52,"value":684},"Analyze token usage",{"type":46,"tag":343,"props":686,"children":687},{},[688,693],{"type":46,"tag":62,"props":689,"children":691},{"className":690},[],[692],{"type":52,"value":525},{"type":52,"value":694}," (Pattern 9)",{"type":46,"tag":316,"props":696,"children":697},{},[698,702,707],{"type":46,"tag":343,"props":699,"children":700},{},[701],{"type":52,"value":401},{"type":46,"tag":343,"props":703,"children":704},{},[705],{"type":52,"value":706},"Detect context issues",{"type":46,"tag":343,"props":708,"children":709},{},[710,716],{"type":46,"tag":62,"props":711,"children":713},{"className":712},[],[714],{"type":52,"value":715},"patterns-context-optimization.md",{"type":52,"value":717}," (Section 5)",{"type":46,"tag":316,"props":719,"children":720},{},[721,725,730],{"type":46,"tag":343,"props":722,"children":723},{},[724],{"type":52,"value":424},{"type":46,"tag":343,"props":726,"children":727},{},[728],{"type":52,"value":729},"Apply optimizations",{"type":46,"tag":343,"props":731,"children":732},{},[733,738],{"type":46,"tag":62,"props":734,"children":736},{"className":735},[],[737],{"type":52,"value":715},{"type":52,"value":739}," (Sections 1-4, 6)",{"type":46,"tag":316,"props":741,"children":742},{},[743,747,752],{"type":46,"tag":343,"props":744,"children":745},{},[746],{"type":52,"value":453},{"type":46,"tag":343,"props":748,"children":749},{},[750],{"type":52,"value":751},"Re-evaluate to measure impact",{"type":46,"tag":343,"props":753,"children":754},{},[755,760],{"type":46,"tag":62,"props":756,"children":758},{"className":757},[],[759],{"type":52,"value":467},{"type":52,"value":761}," (Pattern 6-7)",{"type":46,"tag":296,"props":763,"children":765},{"id":764},"workflow-4-regression-detection",[766],{"type":52,"value":767},"Workflow 4: Regression Detection",{"type":46,"tag":69,"props":769,"children":770},{},[771],{"type":52,"value":772},"For comparing agent versions and finding regressions.",{"type":46,"tag":308,"props":774,"children":775},{},[776,794],{"type":46,"tag":312,"props":777,"children":778},{},[779],{"type":46,"tag":316,"props":780,"children":781},{},[782,786,790],{"type":46,"tag":320,"props":783,"children":784},{},[785],{"type":52,"value":324},{"type":46,"tag":320,"props":787,"children":788},{},[789],{"type":52,"value":329},{"type":46,"tag":320,"props":791,"children":792},{},[793],{"type":52,"value":334},{"type":46,"tag":336,"props":795,"children":796},{},[797,819,841,862,884],{"type":46,"tag":316,"props":798,"children":799},{},[800,804,809],{"type":46,"tag":343,"props":801,"children":802},{},[803],{"type":52,"value":347},{"type":46,"tag":343,"props":805,"children":806},{},[807],{"type":52,"value":808},"Establish baseline",{"type":46,"tag":343,"props":810,"children":811},{},[812,817],{"type":46,"tag":62,"props":813,"children":815},{"className":814},[],[816],{"type":52,"value":467},{"type":52,"value":818}," (Pattern 4: named runs)",{"type":46,"tag":316,"props":820,"children":821},{},[822,826,831],{"type":46,"tag":343,"props":823,"children":824},{},[825],{"type":52,"value":371},{"type":46,"tag":343,"props":827,"children":828},{},[829],{"type":52,"value":830},"Run current version",{"type":46,"tag":343,"props":832,"children":833},{},[834,839],{"type":46,"tag":62,"props":835,"children":837},{"className":836},[],[838],{"type":52,"value":467},{"type":52,"value":840}," (Pattern 1)",{"type":46,"tag":316,"props":842,"children":843},{},[844,848,853],{"type":46,"tag":343,"props":845,"children":846},{},[847],{"type":52,"value":401},{"type":46,"tag":343,"props":849,"children":850},{},[851],{"type":52,"value":852},"Compare metrics",{"type":46,"tag":343,"props":854,"children":855},{},[856,861],{"type":46,"tag":62,"props":857,"children":859},{"className":858},[],[860],{"type":52,"value":467},{"type":52,"value":593},{"type":46,"tag":316,"props":863,"children":864},{},[865,869,874],{"type":46,"tag":343,"props":866,"children":867},{},[868],{"type":52,"value":424},{"type":46,"tag":343,"props":870,"children":871},{},[872],{"type":52,"value":873},"Analyze failing traces",{"type":46,"tag":343,"props":875,"children":876},{},[877,882],{"type":46,"tag":62,"props":878,"children":880},{"className":879},[],[881],{"type":52,"value":525},{"type":52,"value":883}," (Pattern 7)",{"type":46,"tag":316,"props":885,"children":886},{},[887,891,896],{"type":46,"tag":343,"props":888,"children":889},{},[890],{"type":52,"value":453},{"type":46,"tag":343,"props":892,"children":893},{},[894],{"type":52,"value":895},"Debug specific failures",{"type":46,"tag":343,"props":897,"children":898},{},[899,904],{"type":46,"tag":62,"props":900,"children":902},{"className":901},[],[903],{"type":52,"value":525},{"type":52,"value":905}," (Patterns 8-9)",{"type":46,"tag":296,"props":907,"children":909},{"id":908},"workflow-5-custom-scorer-development",[910],{"type":52,"value":911},"Workflow 5: Custom Scorer Development",{"type":46,"tag":69,"props":913,"children":914},{},[915],{"type":52,"value":916},"For creating project-specific evaluation metrics.",{"type":46,"tag":308,"props":918,"children":919},{},[920,938],{"type":46,"tag":312,"props":921,"children":922},{},[923],{"type":46,"tag":316,"props":924,"children":925},{},[926,930,934],{"type":46,"tag":320,"props":927,"children":928},{},[929],{"type":52,"value":324},{"type":46,"tag":320,"props":931,"children":932},{},[933],{"type":52,"value":329},{"type":46,"tag":320,"props":935,"children":936},{},[937],{"type":52,"value":334},{"type":46,"tag":336,"props":939,"children":940},{},[941,963,985,1007],{"type":46,"tag":316,"props":942,"children":943},{},[944,948,953],{"type":46,"tag":343,"props":945,"children":946},{},[947],{"type":52,"value":347},{"type":46,"tag":343,"props":949,"children":950},{},[951],{"type":52,"value":952},"Understand scorer interface",{"type":46,"tag":343,"props":954,"children":955},{},[956,961],{"type":46,"tag":62,"props":957,"children":959},{"className":958},[],[960],{"type":52,"value":393},{"type":52,"value":962}," (Scorer section)",{"type":46,"tag":316,"props":964,"children":965},{},[966,970,975],{"type":46,"tag":343,"props":967,"children":968},{},[969],{"type":52,"value":371},{"type":46,"tag":343,"props":971,"children":972},{},[973],{"type":52,"value":974},"Choose scorer pattern",{"type":46,"tag":343,"props":976,"children":977},{},[978,983],{"type":46,"tag":62,"props":979,"children":981},{"className":980},[],[982],{"type":52,"value":232},{"type":52,"value":984}," (Patterns 4-11)",{"type":46,"tag":316,"props":986,"children":987},{},[988,992,997],{"type":46,"tag":343,"props":989,"children":990},{},[991],{"type":52,"value":401},{"type":46,"tag":343,"props":993,"children":994},{},[995],{"type":52,"value":996},"For multi-agent scorers",{"type":46,"tag":343,"props":998,"children":999},{},[1000,1005],{"type":46,"tag":62,"props":1001,"children":1003},{"className":1002},[],[1004],{"type":52,"value":232},{"type":52,"value":1006}," (Patterns 13-16)",{"type":46,"tag":316,"props":1008,"children":1009},{},[1010,1014,1019],{"type":46,"tag":343,"props":1011,"children":1012},{},[1013],{"type":52,"value":424},{"type":46,"tag":343,"props":1015,"children":1016},{},[1017],{"type":52,"value":1018},"Test with evaluation",{"type":46,"tag":343,"props":1020,"children":1021},{},[1022,1027],{"type":46,"tag":62,"props":1023,"children":1025},{"className":1024},[],[1026],{"type":52,"value":467},{"type":52,"value":840},{"type":46,"tag":296,"props":1029,"children":1031},{"id":1030},"workflow-6-unity-catalog-trace-ingestion-production-monitoring",[1032],{"type":52,"value":1033},"Workflow 6: Unity Catalog Trace Ingestion & Production Monitoring",{"type":46,"tag":69,"props":1035,"children":1036},{},[1037],{"type":52,"value":1038},"For storing traces in Unity Catalog, instrumenting applications, and enabling continuous production monitoring.",{"type":46,"tag":308,"props":1040,"children":1041},{},[1042,1060],{"type":46,"tag":312,"props":1043,"children":1044},{},[1045],{"type":46,"tag":316,"props":1046,"children":1047},{},[1048,1052,1056],{"type":46,"tag":320,"props":1049,"children":1050},{},[1051],{"type":52,"value":324},{"type":46,"tag":320,"props":1053,"children":1054},{},[1055],{"type":52,"value":329},{"type":46,"tag":320,"props":1057,"children":1058},{},[1059],{"type":52,"value":334},{"type":46,"tag":336,"props":1061,"children":1062},{},[1063,1085,1107,1129,1151,1173],{"type":46,"tag":316,"props":1064,"children":1065},{},[1066,1070,1075],{"type":46,"tag":343,"props":1067,"children":1068},{},[1069],{"type":52,"value":347},{"type":46,"tag":343,"props":1071,"children":1072},{},[1073],{"type":52,"value":1074},"Link UC schema to experiment",{"type":46,"tag":343,"props":1076,"children":1077},{},[1078,1083],{"type":46,"tag":62,"props":1079,"children":1081},{"className":1080},[],[1082],{"type":52,"value":167},{"type":52,"value":1084}," (Patterns 1-2)",{"type":46,"tag":316,"props":1086,"children":1087},{},[1088,1092,1097],{"type":46,"tag":343,"props":1089,"children":1090},{},[1091],{"type":52,"value":371},{"type":46,"tag":343,"props":1093,"children":1094},{},[1095],{"type":52,"value":1096},"Set trace destination",{"type":46,"tag":343,"props":1098,"children":1099},{},[1100,1105],{"type":46,"tag":62,"props":1101,"children":1103},{"className":1102},[],[1104],{"type":52,"value":167},{"type":52,"value":1106}," (Patterns 3-4)",{"type":46,"tag":316,"props":1108,"children":1109},{},[1110,1114,1119],{"type":46,"tag":343,"props":1111,"children":1112},{},[1113],{"type":52,"value":401},{"type":46,"tag":343,"props":1115,"children":1116},{},[1117],{"type":52,"value":1118},"Instrument your application",{"type":46,"tag":343,"props":1120,"children":1121},{},[1122,1127],{"type":46,"tag":62,"props":1123,"children":1125},{"className":1124},[],[1126],{"type":52,"value":167},{"type":52,"value":1128}," (Patterns 5-8)",{"type":46,"tag":316,"props":1130,"children":1131},{},[1132,1136,1141],{"type":46,"tag":343,"props":1133,"children":1134},{},[1135],{"type":52,"value":424},{"type":46,"tag":343,"props":1137,"children":1138},{},[1139],{"type":52,"value":1140},"Configure trace sources (Apps\u002FServing\u002FOTEL)",{"type":46,"tag":343,"props":1142,"children":1143},{},[1144,1149],{"type":46,"tag":62,"props":1145,"children":1147},{"className":1146},[],[1148],{"type":52,"value":167},{"type":52,"value":1150}," (Patterns 9-11)",{"type":46,"tag":316,"props":1152,"children":1153},{},[1154,1158,1163],{"type":46,"tag":343,"props":1155,"children":1156},{},[1157],{"type":52,"value":453},{"type":46,"tag":343,"props":1159,"children":1160},{},[1161],{"type":52,"value":1162},"Enable production monitoring",{"type":46,"tag":343,"props":1164,"children":1165},{},[1166,1171],{"type":46,"tag":62,"props":1167,"children":1169},{"className":1168},[],[1170],{"type":52,"value":167},{"type":52,"value":1172}," (Patterns 12-13)",{"type":46,"tag":316,"props":1174,"children":1175},{},[1176,1181,1186],{"type":46,"tag":343,"props":1177,"children":1178},{},[1179],{"type":52,"value":1180},"6",{"type":46,"tag":343,"props":1182,"children":1183},{},[1184],{"type":52,"value":1185},"Query and analyze UC traces",{"type":46,"tag":343,"props":1187,"children":1188},{},[1189,1194],{"type":46,"tag":62,"props":1190,"children":1192},{"className":1191},[],[1193],{"type":52,"value":167},{"type":52,"value":1195}," (Pattern 14)",{"type":46,"tag":296,"props":1197,"children":1199},{"id":1198},"workflow-7-judge-alignment-with-memalign",[1200],{"type":52,"value":1201},"Workflow 7: Judge Alignment with MemAlign",{"type":46,"tag":69,"props":1203,"children":1204},{},[1205],{"type":52,"value":1206},"For aligning an LLM judge to match domain expert preferences. A well-aligned judge improves every downstream use: evaluation accuracy, production monitoring signal, and prompt optimization quality. This workflow is valuable on its own, independent of prompt optimization.",{"type":46,"tag":308,"props":1208,"children":1209},{},[1210,1228],{"type":46,"tag":312,"props":1211,"children":1212},{},[1213],{"type":46,"tag":316,"props":1214,"children":1215},{},[1216,1220,1224],{"type":46,"tag":320,"props":1217,"children":1218},{},[1219],{"type":52,"value":324},{"type":46,"tag":320,"props":1221,"children":1222},{},[1223],{"type":52,"value":329},{"type":46,"tag":320,"props":1225,"children":1226},{},[1227],{"type":52,"value":334},{"type":46,"tag":336,"props":1229,"children":1230},{},[1231,1260,1281,1303,1325,1347],{"type":46,"tag":316,"props":1232,"children":1233},{},[1234,1238,1251],{"type":46,"tag":343,"props":1235,"children":1236},{},[1237],{"type":52,"value":347},{"type":46,"tag":343,"props":1239,"children":1240},{},[1241,1243,1249],{"type":52,"value":1242},"Design base judge with ",{"type":46,"tag":62,"props":1244,"children":1246},{"className":1245},[],[1247],{"type":52,"value":1248},"make_judge",{"type":52,"value":1250}," (any feedback type)",{"type":46,"tag":343,"props":1252,"children":1253},{},[1254,1259],{"type":46,"tag":62,"props":1255,"children":1257},{"className":1256},[],[1258],{"type":52,"value":185},{"type":52,"value":840},{"type":46,"tag":316,"props":1261,"children":1262},{},[1263,1267,1272],{"type":46,"tag":343,"props":1264,"children":1265},{},[1266],{"type":52,"value":371},{"type":46,"tag":343,"props":1268,"children":1269},{},[1270],{"type":52,"value":1271},"Run evaluate(), tag successful traces",{"type":46,"tag":343,"props":1273,"children":1274},{},[1275,1280],{"type":46,"tag":62,"props":1276,"children":1278},{"className":1277},[],[1279],{"type":52,"value":185},{"type":52,"value":615},{"type":46,"tag":316,"props":1282,"children":1283},{},[1284,1288,1293],{"type":46,"tag":343,"props":1285,"children":1286},{},[1287],{"type":52,"value":401},{"type":46,"tag":343,"props":1289,"children":1290},{},[1291],{"type":52,"value":1292},"Build UC dataset + create SME labeling session",{"type":46,"tag":343,"props":1294,"children":1295},{},[1296,1301],{"type":46,"tag":62,"props":1297,"children":1299},{"className":1298},[],[1300],{"type":52,"value":185},{"type":52,"value":1302}," (Pattern 3)",{"type":46,"tag":316,"props":1304,"children":1305},{},[1306,1310,1315],{"type":46,"tag":343,"props":1307,"children":1308},{},[1309],{"type":52,"value":424},{"type":46,"tag":343,"props":1311,"children":1312},{},[1313],{"type":52,"value":1314},"Align judge with MemAlign after labeling completes",{"type":46,"tag":343,"props":1316,"children":1317},{},[1318,1323],{"type":46,"tag":62,"props":1319,"children":1321},{"className":1320},[],[1322],{"type":52,"value":185},{"type":52,"value":1324}," (Pattern 4)",{"type":46,"tag":316,"props":1326,"children":1327},{},[1328,1332,1337],{"type":46,"tag":343,"props":1329,"children":1330},{},[1331],{"type":52,"value":453},{"type":46,"tag":343,"props":1333,"children":1334},{},[1335],{"type":52,"value":1336},"Register aligned judge to experiment",{"type":46,"tag":343,"props":1338,"children":1339},{},[1340,1345],{"type":46,"tag":62,"props":1341,"children":1343},{"className":1342},[],[1344],{"type":52,"value":185},{"type":52,"value":1346}," (Pattern 5)",{"type":46,"tag":316,"props":1348,"children":1349},{},[1350,1354,1359],{"type":46,"tag":343,"props":1351,"children":1352},{},[1353],{"type":52,"value":1180},{"type":46,"tag":343,"props":1355,"children":1356},{},[1357],{"type":52,"value":1358},"Re-evaluate with aligned judge (baseline)",{"type":46,"tag":343,"props":1360,"children":1361},{},[1362,1367],{"type":46,"tag":62,"props":1363,"children":1365},{"className":1364},[],[1366],{"type":52,"value":185},{"type":52,"value":1368}," (Pattern 6)",{"type":46,"tag":296,"props":1370,"children":1372},{"id":1371},"workflow-8-automated-prompt-optimization-with-gepa",[1373],{"type":52,"value":1374},"Workflow 8: Automated Prompt Optimization with GEPA",{"type":46,"tag":69,"props":1376,"children":1377},{},[1378,1380,1385,1387,1392],{"type":52,"value":1379},"For automatically improving a registered system prompt using ",{"type":46,"tag":62,"props":1381,"children":1383},{"className":1382},[],[1384],{"type":52,"value":198},{"type":52,"value":1386},". Works with any scorer, but paired with an aligned judge (Workflow 7) gives the most domain-accurate signal. For the full end-to-end loop combining alignment and optimization, see ",{"type":46,"tag":62,"props":1388,"children":1390},{"className":1389},[],[1391],{"type":52,"value":361},{"type":52,"value":1393}," Journey 10.",{"type":46,"tag":308,"props":1395,"children":1396},{},[1397,1415],{"type":46,"tag":312,"props":1398,"children":1399},{},[1400],{"type":46,"tag":316,"props":1401,"children":1402},{},[1403,1407,1411],{"type":46,"tag":320,"props":1404,"children":1405},{},[1406],{"type":52,"value":324},{"type":46,"tag":320,"props":1408,"children":1409},{},[1410],{"type":52,"value":329},{"type":46,"tag":320,"props":1412,"children":1413},{},[1414],{"type":52,"value":334},{"type":46,"tag":336,"props":1416,"children":1417},{},[1418,1439,1460],{"type":46,"tag":316,"props":1419,"children":1420},{},[1421,1425,1430],{"type":46,"tag":343,"props":1422,"children":1423},{},[1424],{"type":52,"value":347},{"type":46,"tag":343,"props":1426,"children":1427},{},[1428],{"type":52,"value":1429},"Build optimization dataset (inputs + expectations)",{"type":46,"tag":343,"props":1431,"children":1432},{},[1433,1438],{"type":46,"tag":62,"props":1434,"children":1436},{"className":1435},[],[1437],{"type":52,"value":208},{"type":52,"value":840},{"type":46,"tag":316,"props":1440,"children":1441},{},[1442,1446,1451],{"type":46,"tag":343,"props":1443,"children":1444},{},[1445],{"type":52,"value":371},{"type":46,"tag":343,"props":1447,"children":1448},{},[1449],{"type":52,"value":1450},"Run optimize_prompts() with GEPA + scorer",{"type":46,"tag":343,"props":1452,"children":1453},{},[1454,1459],{"type":46,"tag":62,"props":1455,"children":1457},{"className":1456},[],[1458],{"type":52,"value":208},{"type":52,"value":615},{"type":46,"tag":316,"props":1461,"children":1462},{},[1463,1467,1472],{"type":46,"tag":343,"props":1464,"children":1465},{},[1466],{"type":52,"value":401},{"type":46,"tag":343,"props":1468,"children":1469},{},[1470],{"type":52,"value":1471},"Register new version, promote conditionally",{"type":46,"tag":343,"props":1473,"children":1474},{},[1475,1480],{"type":46,"tag":62,"props":1476,"children":1478},{"className":1477},[],[1479],{"type":52,"value":208},{"type":52,"value":1302},{"type":46,"tag":55,"props":1482,"children":1484},{"id":1483},"reference-files-quick-lookup",[1485],{"type":52,"value":1486},"Reference Files Quick Lookup",{"type":46,"tag":308,"props":1488,"children":1489},{},[1490,1511],{"type":46,"tag":312,"props":1491,"children":1492},{},[1493],{"type":46,"tag":316,"props":1494,"children":1495},{},[1496,1501,1506],{"type":46,"tag":320,"props":1497,"children":1498},{},[1499],{"type":52,"value":1500},"Reference",{"type":46,"tag":320,"props":1502,"children":1503},{},[1504],{"type":52,"value":1505},"Purpose",{"type":46,"tag":320,"props":1507,"children":1508},{},[1509],{"type":52,"value":1510},"When to Read",{"type":46,"tag":336,"props":1512,"children":1513},{},[1514,1540,1561,1582,1603,1624,1645,1666,1687,1708,1729],{"type":46,"tag":316,"props":1515,"children":1516},{},[1517,1525,1530],{"type":46,"tag":343,"props":1518,"children":1519},{},[1520],{"type":46,"tag":62,"props":1521,"children":1523},{"className":1522},[],[1524],{"type":52,"value":385},{"type":46,"tag":343,"props":1526,"children":1527},{},[1528],{"type":52,"value":1529},"Common mistakes",{"type":46,"tag":343,"props":1531,"children":1532},{},[1533,1538],{"type":46,"tag":140,"props":1534,"children":1535},{},[1536],{"type":52,"value":1537},"Always read first",{"type":52,"value":1539}," before writing code",{"type":46,"tag":316,"props":1541,"children":1542},{},[1543,1551,1556],{"type":46,"tag":343,"props":1544,"children":1545},{},[1546],{"type":46,"tag":62,"props":1547,"children":1549},{"className":1548},[],[1550],{"type":52,"value":393},{"type":46,"tag":343,"props":1552,"children":1553},{},[1554],{"type":52,"value":1555},"API signatures, schemas",{"type":46,"tag":343,"props":1557,"children":1558},{},[1559],{"type":52,"value":1560},"When writing any evaluation code",{"type":46,"tag":316,"props":1562,"children":1563},{},[1564,1572,1577],{"type":46,"tag":343,"props":1565,"children":1566},{},[1567],{"type":46,"tag":62,"props":1568,"children":1570},{"className":1569},[],[1571],{"type":52,"value":467},{"type":46,"tag":343,"props":1573,"children":1574},{},[1575],{"type":52,"value":1576},"Running evals, comparing",{"type":46,"tag":343,"props":1578,"children":1579},{},[1580],{"type":52,"value":1581},"When executing evaluations",{"type":46,"tag":316,"props":1583,"children":1584},{},[1585,1593,1598],{"type":46,"tag":343,"props":1586,"children":1587},{},[1588],{"type":46,"tag":62,"props":1589,"children":1591},{"className":1590},[],[1592],{"type":52,"value":232},{"type":46,"tag":343,"props":1594,"children":1595},{},[1596],{"type":52,"value":1597},"Custom scorer creation",{"type":46,"tag":343,"props":1599,"children":1600},{},[1601],{"type":52,"value":1602},"When built-in scorers aren't enough",{"type":46,"tag":316,"props":1604,"children":1605},{},[1606,1614,1619],{"type":46,"tag":343,"props":1607,"children":1608},{},[1609],{"type":46,"tag":62,"props":1610,"children":1612},{"className":1611},[],[1613],{"type":52,"value":225},{"type":46,"tag":343,"props":1615,"children":1616},{},[1617],{"type":52,"value":1618},"Dataset building",{"type":46,"tag":343,"props":1620,"children":1621},{},[1622],{"type":52,"value":1623},"When preparing evaluation data",{"type":46,"tag":316,"props":1625,"children":1626},{},[1627,1635,1640],{"type":46,"tag":343,"props":1628,"children":1629},{},[1630],{"type":46,"tag":62,"props":1631,"children":1633},{"className":1632},[],[1634],{"type":52,"value":525},{"type":46,"tag":343,"props":1636,"children":1637},{},[1638],{"type":52,"value":1639},"Trace debugging",{"type":46,"tag":343,"props":1641,"children":1642},{},[1643],{"type":52,"value":1644},"When analyzing agent behavior",{"type":46,"tag":316,"props":1646,"children":1647},{},[1648,1656,1661],{"type":46,"tag":343,"props":1649,"children":1650},{},[1651],{"type":46,"tag":62,"props":1652,"children":1654},{"className":1653},[],[1655],{"type":52,"value":715},{"type":46,"tag":343,"props":1657,"children":1658},{},[1659],{"type":52,"value":1660},"Token\u002Flatency fixes",{"type":46,"tag":343,"props":1662,"children":1663},{},[1664],{"type":52,"value":1665},"When agent is slow or expensive",{"type":46,"tag":316,"props":1667,"children":1668},{},[1669,1677,1682],{"type":46,"tag":343,"props":1670,"children":1671},{},[1672],{"type":46,"tag":62,"props":1673,"children":1675},{"className":1674},[],[1676],{"type":52,"value":167},{"type":46,"tag":343,"props":1678,"children":1679},{},[1680],{"type":52,"value":1681},"UC trace setup, monitoring",{"type":46,"tag":343,"props":1683,"children":1684},{},[1685],{"type":52,"value":1686},"When setting up trace storage or production monitoring",{"type":46,"tag":316,"props":1688,"children":1689},{},[1690,1698,1703],{"type":46,"tag":343,"props":1691,"children":1692},{},[1693],{"type":46,"tag":62,"props":1694,"children":1696},{"className":1695},[],[1697],{"type":52,"value":185},{"type":46,"tag":343,"props":1699,"children":1700},{},[1701],{"type":52,"value":1702},"MemAlign judge alignment, labeling sessions, SME feedback",{"type":46,"tag":343,"props":1704,"children":1705},{},[1706],{"type":52,"value":1707},"When aligning judges to domain expert preferences",{"type":46,"tag":316,"props":1709,"children":1710},{},[1711,1719,1724],{"type":46,"tag":343,"props":1712,"children":1713},{},[1714],{"type":46,"tag":62,"props":1715,"children":1717},{"className":1716},[],[1718],{"type":52,"value":208},{"type":46,"tag":343,"props":1720,"children":1721},{},[1722],{"type":52,"value":1723},"GEPA optimization: build dataset, optimize_prompts(), promote",{"type":46,"tag":343,"props":1725,"children":1726},{},[1727],{"type":52,"value":1728},"When running automated prompt improvement",{"type":46,"tag":316,"props":1730,"children":1731},{},[1732,1740,1745],{"type":46,"tag":343,"props":1733,"children":1734},{},[1735],{"type":46,"tag":62,"props":1736,"children":1738},{"className":1737},[],[1739],{"type":52,"value":361},{"type":46,"tag":343,"props":1741,"children":1742},{},[1743],{"type":52,"value":1744},"High-level workflows, full domain-expert optimization loop",{"type":46,"tag":343,"props":1746,"children":1747},{},[1748],{"type":52,"value":1749},"When starting a new evaluation project or running the full align + optimize cycle",{"type":46,"tag":55,"props":1751,"children":1753},{"id":1752},"critical-api-facts",[1754],{"type":52,"value":1755},"Critical API Facts",{"type":46,"tag":148,"props":1757,"children":1758},{},[1759,1784,1801,1819,1845,1886,1896,1922,1940],{"type":46,"tag":152,"props":1760,"children":1761},{},[1762,1767,1769,1774,1776,1782],{"type":46,"tag":140,"props":1763,"children":1764},{},[1765],{"type":52,"value":1766},"Use:",{"type":52,"value":1768}," ",{"type":46,"tag":62,"props":1770,"children":1772},{"className":1771},[],[1773],{"type":52,"value":131},{"type":52,"value":1775}," (NOT ",{"type":46,"tag":62,"props":1777,"children":1779},{"className":1778},[],[1780],{"type":52,"value":1781},"mlflow.evaluate()",{"type":52,"value":1783},")",{"type":46,"tag":152,"props":1785,"children":1786},{},[1787,1792,1793,1799],{"type":46,"tag":140,"props":1788,"children":1789},{},[1790],{"type":52,"value":1791},"Data format:",{"type":52,"value":1768},{"type":46,"tag":62,"props":1794,"children":1796},{"className":1795},[],[1797],{"type":52,"value":1798},"{\"inputs\": {\"query\": \"...\"}}",{"type":52,"value":1800}," (nested structure required)",{"type":46,"tag":152,"props":1802,"children":1803},{},[1804,1809,1811,1817],{"type":46,"tag":140,"props":1805,"children":1806},{},[1807],{"type":52,"value":1808},"predict_fn:",{"type":52,"value":1810}," Receives ",{"type":46,"tag":62,"props":1812,"children":1814},{"className":1813},[],[1815],{"type":52,"value":1816},"**unpacked kwargs",{"type":52,"value":1818}," (not a dict)",{"type":46,"tag":152,"props":1820,"children":1821},{},[1822,1827,1829,1835,1837,1843],{"type":46,"tag":140,"props":1823,"children":1824},{},[1825],{"type":52,"value":1826},"MemAlign:",{"type":52,"value":1828}," Scorer-agnostic (works with any ",{"type":46,"tag":62,"props":1830,"children":1832},{"className":1831},[],[1833],{"type":52,"value":1834},"feedback_value_type",{"type":52,"value":1836}," -- float, bool, categorical); token-heavy on the embedding model so set ",{"type":46,"tag":62,"props":1838,"children":1840},{"className":1839},[],[1841],{"type":52,"value":1842},"embedding_model",{"type":52,"value":1844}," explicitly",{"type":46,"tag":152,"props":1846,"children":1847},{},[1848,1853,1855,1861,1863,1868,1870,1876,1878,1884],{"type":46,"tag":140,"props":1849,"children":1850},{},[1851],{"type":52,"value":1852},"Label schema name matching:",{"type":52,"value":1854}," The label schema ",{"type":46,"tag":62,"props":1856,"children":1858},{"className":1857},[],[1859],{"type":52,"value":1860},"name",{"type":52,"value":1862}," in the labeling session MUST match the judge ",{"type":46,"tag":62,"props":1864,"children":1866},{"className":1865},[],[1867],{"type":52,"value":1860},{"type":52,"value":1869}," used in ",{"type":46,"tag":62,"props":1871,"children":1873},{"className":1872},[],[1874],{"type":52,"value":1875},"evaluate()",{"type":52,"value":1877}," for ",{"type":46,"tag":62,"props":1879,"children":1881},{"className":1880},[],[1882],{"type":52,"value":1883},"align()",{"type":52,"value":1885}," to pair scores",{"type":46,"tag":152,"props":1887,"children":1888},{},[1889,1894],{"type":46,"tag":140,"props":1890,"children":1891},{},[1892],{"type":52,"value":1893},"Aligned judge scores:",{"type":52,"value":1895}," May be lower than unaligned judge scores -- this is expected and means the judge is now more accurate, not that the agent regressed",{"type":46,"tag":152,"props":1897,"children":1898},{},[1899,1904,1906,1912,1914,1920],{"type":46,"tag":140,"props":1900,"children":1901},{},[1902],{"type":52,"value":1903},"GEPA optimization dataset:",{"type":52,"value":1905}," Must have both ",{"type":46,"tag":62,"props":1907,"children":1909},{"className":1908},[],[1910],{"type":52,"value":1911},"inputs",{"type":52,"value":1913}," AND ",{"type":46,"tag":62,"props":1915,"children":1917},{"className":1916},[],[1918],{"type":52,"value":1919},"expectations",{"type":52,"value":1921}," per record (different from eval dataset)",{"type":46,"tag":152,"props":1923,"children":1924},{},[1925,1930,1932,1938],{"type":46,"tag":140,"props":1926,"children":1927},{},[1928],{"type":52,"value":1929},"Episodic memory:",{"type":52,"value":1931}," Lazily loaded -- ",{"type":46,"tag":62,"props":1933,"children":1935},{"className":1934},[],[1936],{"type":52,"value":1937},"get_scorer()",{"type":52,"value":1939}," results won't show episodic memory on print until the judge is first used",{"type":46,"tag":152,"props":1941,"children":1942},{},[1943,1948],{"type":46,"tag":140,"props":1944,"children":1945},{},[1946],{"type":52,"value":1947},"optimize_prompts:",{"type":52,"value":1949}," Requires MLflow >= 3.5.0",{"type":46,"tag":69,"props":1951,"children":1952},{},[1953,1955,1960],{"type":52,"value":1954},"See ",{"type":46,"tag":62,"props":1956,"children":1958},{"className":1957},[],[1959],{"type":52,"value":385},{"type":52,"value":1961}," for complete list.",{"type":46,"tag":55,"props":1963,"children":1965},{"id":1964},"related-skills",[1966],{"type":52,"value":1967},"Related Skills",{"type":46,"tag":148,"props":1969,"children":1970},{},[1971,1985,1995,2009,2023],{"type":46,"tag":152,"props":1972,"children":1973},{},[1974,1983],{"type":46,"tag":140,"props":1975,"children":1976},{},[1977],{"type":46,"tag":82,"props":1978,"children":1980},{"href":1979},"..\u002Fdatabricks-docs\u002FSKILL.md",[1981],{"type":52,"value":1982},"databricks-docs",{"type":52,"value":1984}," - General Databricks documentation reference",{"type":46,"tag":152,"props":1986,"children":1987},{},[1988,1993],{"type":46,"tag":140,"props":1989,"children":1990},{},[1991],{"type":52,"value":1992},"databricks-model-serving",{"type":52,"value":1994}," - Deploying models and agents to serving endpoints",{"type":46,"tag":152,"props":1996,"children":1997},{},[1998,2007],{"type":46,"tag":140,"props":1999,"children":2000},{},[2001],{"type":46,"tag":82,"props":2002,"children":2004},{"href":2003},"..\u002Fdatabricks-agent-bricks\u002FSKILL.md",[2005],{"type":52,"value":2006},"databricks-agent-bricks",{"type":52,"value":2008}," - Building agents that can be evaluated with this skill",{"type":46,"tag":152,"props":2010,"children":2011},{},[2012,2021],{"type":46,"tag":140,"props":2013,"children":2014},{},[2015],{"type":46,"tag":82,"props":2016,"children":2018},{"href":2017},"..\u002Fdatabricks-python-sdk\u002FSKILL.md",[2019],{"type":52,"value":2020},"databricks-python-sdk",{"type":52,"value":2022}," - SDK patterns used alongside MLflow APIs",{"type":46,"tag":152,"props":2024,"children":2025},{},[2026,2035],{"type":46,"tag":140,"props":2027,"children":2028},{},[2029],{"type":46,"tag":82,"props":2030,"children":2032},{"href":2031},"..\u002Fdatabricks-unity-catalog\u002FSKILL.md",[2033],{"type":52,"value":2034},"databricks-unity-catalog",{"type":52,"value":2036}," - Unity Catalog tables for managed evaluation datasets",{"items":2038,"total":2153},[2039,2053,2065,2082,2099,2119,2130],{"slug":2006,"name":2006,"fn":2040,"description":2041,"org":2042,"tags":2043,"stars":26,"repoUrl":27,"updatedAt":2052},"create Databricks Agent Bricks","Create Agent Bricks: Knowledge Assistants (KA) for document Q&A and Supervisor Agents for multi-agent orchestration (MAS).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2044,2045,2046,2049],{"name":24,"slug":25,"type":15},{"name":9,"slug":8,"type":15},{"name":2047,"slug":2048,"type":15},"Knowledge Management","knowledge-management",{"name":2050,"slug":2051,"type":15},"Multi-Agent","multi-agent","2026-07-15T05:41:38.548954",{"slug":2054,"name":2054,"fn":2055,"description":2056,"org":2057,"tags":2058,"stars":26,"repoUrl":27,"updatedAt":2064},"databricks-ai-functions","use Databricks built-in AI functions","Use Databricks built-in AI Functions (ai_classify, ai_extract, ai_summarize, ai_mask, ai_translate, ai_fix_grammar, ai_gen, ai_analyze_sentiment, ai_similarity, ai_parse_document, ai_prep_search, ai_query, ai_forecast) to add AI capabilities directly to SQL and PySpark pipelines without managing model endpoints. Also covers document parsing and building custom RAG pipelines (parse → prep_search → index → query).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2059,2062,2063],{"name":2060,"slug":2061,"type":15},"Data Analysis","data-analysis",{"name":9,"slug":8,"type":15},{"name":13,"slug":14,"type":15},"2026-07-31T05:53:33.562077",{"slug":2066,"name":2066,"fn":2067,"description":2068,"org":2069,"tags":2070,"stars":26,"repoUrl":27,"updatedAt":2081},"databricks-ai-runtime","submit and manage Databricks GPU workloads","Databricks AI Runtime (`air`) CLI — the command-line tool for submitting and managing GPU training workloads on Databricks serverless compute. Use for: running `air` workloads, custom Docker image setup, environment configuration, and troubleshooting `air` jobs.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2071,2074,2075,2078],{"name":2072,"slug":2073,"type":15},"CLI","cli",{"name":9,"slug":8,"type":15},{"name":2076,"slug":2077,"type":15},"Docker","docker",{"name":2079,"slug":2080,"type":15},"Engineering","engineering","2026-07-12T08:04:55.843982",{"slug":2083,"name":2083,"fn":2084,"description":2085,"org":2086,"tags":2087,"stars":26,"repoUrl":27,"updatedAt":2098},"databricks-aibi-dashboards","create Databricks AI\u002FBI dashboards","Create Databricks AI\u002FBI dashboards. Must use when creating, updating, or deploying Lakeview dashboards as Databricks Dashboard have a unique json structure. CRITICAL: You MUST test ALL SQL queries via CLI BEFORE deploying. Follow guidelines strictly.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2088,2091,2094,2097],{"name":2089,"slug":2090,"type":15},"Analytics","analytics",{"name":2092,"slug":2093,"type":15},"Dashboards","dashboards",{"name":2095,"slug":2096,"type":15},"Data Visualization","data-visualization",{"name":9,"slug":8,"type":15},"2026-07-12T08:04:25.314591",{"slug":2100,"name":2100,"fn":2101,"description":2102,"org":2103,"tags":2104,"stars":26,"repoUrl":27,"updatedAt":2118},"databricks-app-design","design UX for Databricks AppKit applications","Design the UX of custom-code Databricks Apps (AppKit\u002FReact) data screens — KPI\u002Foverview pages, reports, charts, tables, and Genie\u002Fchat data assistants — mapped to concrete AppKit components. Use when BUILDING or reviewing the UI of an AppKit\u002FReact app that displays data or answers data questions: choosing genre, layout, charts, KPIs, semantic color, required states (loading\u002Fempty\u002Ferror), IBCS notation, and AI-result trust (showing generated SQL\u002Fsources for Genie\u002Fchat). A plain \"create a dashboard\" request means a managed AI\u002FBI (Lakeview) dashboard → use databricks-aibi-dashboards, NOT this skill. Also NOT for non-data frontend (forms, settings, auth, marketing) or scaffolding\u002Fbuild\u002Fdeploy (→ databricks-apps). Complements databricks-apps; use it alongside whenever a custom app has a chart, table, KPI, report, or Genie\u002Fchat\u002FAI surface.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2105,2106,2109,2112,2115],{"name":9,"slug":8,"type":15},{"name":2107,"slug":2108,"type":15},"Design","design",{"name":2110,"slug":2111,"type":15},"Frontend","frontend",{"name":2113,"slug":2114,"type":15},"React","react",{"name":2116,"slug":2117,"type":15},"UI Components","ui-components","2026-07-12T08:04:02.02398",{"slug":2120,"name":2120,"fn":2121,"description":2122,"org":2123,"tags":2124,"stars":26,"repoUrl":27,"updatedAt":2129},"databricks-apps","build applications on Databricks Apps","Build apps on Databricks Apps platform. Use when asked to create data apps, analytics tools, or custom interactive visualizations. A plain \"create a dashboard\" request means a managed AI\u002FBI (Lakeview) dashboard → use databricks-aibi-dashboards, not this skill. Evaluates data access patterns (analytics vs Lakebase synced tables) before scaffolding. Invoke BEFORE starting implementation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2125,2126,2127,2128],{"name":2089,"slug":2090,"type":15},{"name":2092,"slug":2093,"type":15},{"name":2060,"slug":2061,"type":15},{"name":9,"slug":8,"type":15},"2026-07-12T08:03:59.061458",{"slug":2131,"name":2131,"fn":2132,"description":2133,"org":2134,"tags":2135,"stars":26,"repoUrl":27,"updatedAt":2152},"databricks-apps-python","build Python backends for Databricks Apps","Python backend for Databricks Apps — FastAPI (default), Flask, Dash, Streamlit, Gradio, Reflex. **Default for a new Databricks App is `databricks-apps` (AppKit — Node\u002FTypeScript\u002FReact) — reach for it first.** Use this skill only when the user asks for a Python backend, extends an existing Python app, or the team is Python-only. Covers OAuth auth, app resources, SQL warehouse and Lakebase connectivity, foundation-model \u002F Vector Search \u002F model-serving APIs (via `databricks-python-sdk`), and deployment via CLI or DABs.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2136,2137,2140,2143,2146,2149],{"name":9,"slug":8,"type":15},{"name":2138,"slug":2139,"type":15},"FastAPI","fastapi",{"name":2141,"slug":2142,"type":15},"Flask","flask",{"name":2144,"slug":2145,"type":15},"Gradio","gradio",{"name":2147,"slug":2148,"type":15},"Python","python",{"name":2150,"slug":2151,"type":15},"Streamlit","streamlit","2026-07-12T08:04:10.970845",31,{"items":2155,"total":2153},[2156,2163,2169,2176,2183,2191,2198,2207,2218,2235,2250,2263],{"slug":2006,"name":2006,"fn":2040,"description":2041,"org":2157,"tags":2158,"stars":26,"repoUrl":27,"updatedAt":2052},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2159,2160,2161,2162],{"name":24,"slug":25,"type":15},{"name":9,"slug":8,"type":15},{"name":2047,"slug":2048,"type":15},{"name":2050,"slug":2051,"type":15},{"slug":2054,"name":2054,"fn":2055,"description":2056,"org":2164,"tags":2165,"stars":26,"repoUrl":27,"updatedAt":2064},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2166,2167,2168],{"name":2060,"slug":2061,"type":15},{"name":9,"slug":8,"type":15},{"name":13,"slug":14,"type":15},{"slug":2066,"name":2066,"fn":2067,"description":2068,"org":2170,"tags":2171,"stars":26,"repoUrl":27,"updatedAt":2081},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2172,2173,2174,2175],{"name":2072,"slug":2073,"type":15},{"name":9,"slug":8,"type":15},{"name":2076,"slug":2077,"type":15},{"name":2079,"slug":2080,"type":15},{"slug":2083,"name":2083,"fn":2084,"description":2085,"org":2177,"tags":2178,"stars":26,"repoUrl":27,"updatedAt":2098},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2179,2180,2181,2182],{"name":2089,"slug":2090,"type":15},{"name":2092,"slug":2093,"type":15},{"name":2095,"slug":2096,"type":15},{"name":9,"slug":8,"type":15},{"slug":2100,"name":2100,"fn":2101,"description":2102,"org":2184,"tags":2185,"stars":26,"repoUrl":27,"updatedAt":2118},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2186,2187,2188,2189,2190],{"name":9,"slug":8,"type":15},{"name":2107,"slug":2108,"type":15},{"name":2110,"slug":2111,"type":15},{"name":2113,"slug":2114,"type":15},{"name":2116,"slug":2117,"type":15},{"slug":2120,"name":2120,"fn":2121,"description":2122,"org":2192,"tags":2193,"stars":26,"repoUrl":27,"updatedAt":2129},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2194,2195,2196,2197],{"name":2089,"slug":2090,"type":15},{"name":2092,"slug":2093,"type":15},{"name":2060,"slug":2061,"type":15},{"name":9,"slug":8,"type":15},{"slug":2131,"name":2131,"fn":2132,"description":2133,"org":2199,"tags":2200,"stars":26,"repoUrl":27,"updatedAt":2152},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2201,2202,2203,2204,2205,2206],{"name":9,"slug":8,"type":15},{"name":2138,"slug":2139,"type":15},{"name":2141,"slug":2142,"type":15},{"name":2144,"slug":2145,"type":15},{"name":2147,"slug":2148,"type":15},{"name":2150,"slug":2151,"type":15},{"slug":41,"name":41,"fn":2208,"description":2209,"org":2210,"tags":2211,"stars":26,"repoUrl":27,"updatedAt":2217},"configure Databricks CLI and authentication","Databricks CLI operations and the parent\u002Fentry-point skill for Databricks CLI use: authentication, profile selection, and bundles. Load this first for CLI, auth, profile, and bundle tasks, then load the matching product skill. For finding or exploring data, answering questions about the data, or generating SQL, load the databricks-data-discovery skill (it routes to Genie One). Contains up-to-date guidelines for Databricks-related CLI tasks.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2212,2215,2216],{"name":2213,"slug":2214,"type":15},"Authentication","authentication",{"name":2072,"slug":2073,"type":15},{"name":9,"slug":8,"type":15},"2026-07-18T05:11:05.45506",{"slug":2219,"name":2219,"fn":2220,"description":2221,"org":2222,"tags":2223,"stars":26,"repoUrl":27,"updatedAt":2234},"databricks-dabs","manage Databricks Declarative Automation Bundles","Create, configure, validate, deploy, run, and manage Declarative Automation Bundles (DABs, formerly Databricks Asset Bundles). Use when working with Databricks resources via DABs including dashboards, jobs, pipelines, alerts, volumes, and apps.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2224,2227,2230,2231],{"name":2225,"slug":2226,"type":15},"Automation","automation",{"name":2228,"slug":2229,"type":15},"Configuration","configuration",{"name":9,"slug":8,"type":15},{"name":2232,"slug":2233,"type":15},"Deployment","deployment","2026-07-15T05:41:35.930355",{"slug":2236,"name":2236,"fn":2237,"description":2238,"org":2239,"tags":2240,"stars":26,"repoUrl":27,"updatedAt":2249},"databricks-data-discovery","discover and query Databricks data","Discover, explore, and query Databricks data via Genie — the CLI equivalent of the Genie One MCP. MUST be invoked whenever the user asks to find or locate data ('what tables are in X', 'where does X live', 'which catalog\u002Fschema has Y'), answer a natural-language question about the data, or write a SQL query.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2241,2242,2245,2246],{"name":2060,"slug":2061,"type":15},{"name":2243,"slug":2244,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":2247,"slug":2248,"type":15},"SQL","sql","2026-07-31T05:53:32.561877",{"slug":2251,"name":2251,"fn":2252,"description":2253,"org":2254,"tags":2255,"stars":26,"repoUrl":27,"updatedAt":2262},"databricks-dbsql","query and script Databricks SQL warehouses","Databricks SQL (DBSQL) advanced features and SQL warehouse capabilities. This skill MUST be invoked when the user mentions: \"DBSQL\", \"Databricks SQL\", \"SQL warehouse\", \"SQL scripting\", \"stored procedure\", \"CALL procedure\", \"materialized view\", \"CREATE MATERIALIZED VIEW\", \"pipe syntax\", \"|>\", \"geospatial\", \"H3\", \"ST_\", \"spatial SQL\", \"collation\", \"COLLATE\", \"ai_query\", \"ai_classify\", \"ai_extract\", \"ai_gen\", \"AI function\", \"http_request\", \"remote_query\", \"read_files\", \"Lakehouse Federation\", \"recursive CTE\", \"WITH RECURSIVE\", \"multi-statement transaction\", \"temp table\", \"temporary view\", \"pipe operator\". SHOULD also invoke when the user asks about SQL best practices, data modeling patterns, or advanced SQL features on Databricks.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2256,2257,2260,2261],{"name":2060,"slug":2061,"type":15},{"name":2258,"slug":2259,"type":15},"Database","database",{"name":9,"slug":8,"type":15},{"name":2247,"slug":2248,"type":15},"2026-07-12T08:04:08.678282",{"slug":1982,"name":1982,"fn":2264,"description":2265,"org":2266,"tags":2267,"stars":26,"repoUrl":27,"updatedAt":2274},"search Databricks documentation","Databricks documentation reference via llms.txt index. Use when other skills do not cover a topic, looking up unfamiliar Databricks features, or needing authoritative docs on APIs, configurations, or platform capabilities.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[2268,2269,2272],{"name":9,"slug":8,"type":15},{"name":2270,"slug":2271,"type":15},"Documentation","documentation",{"name":1500,"slug":2273,"type":15},"reference","2026-07-15T05:41:34.697746"]