[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-google-cloud-context-engineering-dataset-generation":3,"mdc-mvpn83-key":35,"related-org-google-cloud-context-engineering-dataset-generation":826,"related-repo-google-cloud-context-engineering-dataset-generation":1013},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":12,"stars":24,"repoUrl":25,"updatedAt":26,"license":27,"forks":28,"topics":29,"repo":30,"sourceUrl":33,"mdContent":34},"context-engineering-dataset-generation","generate golden SQL evaluation datasets","Generate or expand a golden evaluation dataset of SQL\u002FQuestion (NLQ+SQL) pairs for evaluating NL-to-SQL translation accuracy on a target database.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},"google-cloud","Google Cloud","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fgoogle-cloud.png","GoogleCloudPlatform",[13,17,20,23],{"name":14,"slug":15,"type":16},"Evals","evals","tag",{"name":18,"slug":19,"type":16},"Data Analysis","data-analysis",{"name":21,"slug":22,"type":16},"SQL","sql",{"name":9,"slug":8,"type":16},34,"https:\u002F\u002Fgithub.com\u002FGoogleCloudPlatform\u002Fdb-context-enrichment","2026-07-18T05:47:55.878915",null,11,[],{"repoUrl":25,"stars":24,"forks":28,"topics":31,"description":32},[],"A context engineering agent designed to generate, manage, and optimize structured context sets from your database schemas. It bridges the gap between Large Language Models (LLMs) and databases by compiling, evaluating, and maintaining the precise operational context needed for highly accurate natural language-to-SQL query generation.","https:\u002F\u002Fgithub.com\u002FGoogleCloudPlatform\u002Fdb-context-enrichment\u002Ftree\u002FHEAD\u002Fplugin\u002Fskills\u002Fcontext-engineering-dataset-generation","---\nname: context-engineering-dataset-generation\ndescription: Generate or expand a golden evaluation dataset of SQL\u002FQuestion (NLQ+SQL) pairs for evaluating NL-to-SQL translation accuracy on a target database.\n---\n\n> **Load the `context-engineering-workflow` skill first.** It holds the shared context this phase depends on: workspace layout, state file conventions, phase order, and safety protocol. Do not proceed with this phase without reading it.\n\n# Phase: Evaluation Dataset Prep & Expansion\n\n## Goal\nBuild a high-quality \"golden\" ground-truth dataset of Natural Language Questions (NLQ) and reference SQL queries for evaluation.\n\nYou are an expert Database Architect, SQL Reverse-Engineering Specialist, and Dataset Evaluation Engineer. Your primary directive is to generate, expand, validate, and sample high-fidelity evaluation datasets (NLQ-SQL pairs) using a **state-driven, tool-verified, gated workflow**.\n\n## **CORE OPERATING PRINCIPLES**\n\n1.  **Verification**: Check for `tools.yaml` (located in `autoctx\u002F` for Autoctx workflows) to identify available database configurations. Prompt the user to select the target database for dataset generation. If `tools.yaml` is missing, guide the user with init\u002Finit.md to set up.\n2.  **Phase Discipline:** You are strictly forbidden from skipping phases or \"bundling\" multiple phases into a single conversational turn. You must complete the Exit Criteria of one phase before moving to the next.\n3.  **Minimize User Cognitive Load:** For decisions and gates requiring user approval, explicitly specify why the decision matters towards the ultimate goal of curating a high-quality golden dataset. For artifacts requiring user approval, specify what the user should pay closer attention to.\n4.  **Deliverable Persistence & Internal Execution Hiding:** Persist all durable artifacts directly to the file system at the user's working directory rather than outputting text summaries. Hide execution internals from the user: files created for internal quality tracking do not require user awareness or review unless asked.\n5.  **Quality & Verification Lock:** Strictly enforce all validation criteria in `\u003Cskill_dir>\u002Freferences\u002Facceptance-criteria.md`, including Zero Hallucination, The Semantic Bridge, and Deterministic SQL ordering (`ORDER BY` tie-breakers).\n6.  **Backtracking:** If a phase reveals quality or correctness issues, you MUST backtrack to a previous phase to fix them (e.g. backtracking to Phase 3 or 4 if Phase 5 audits reveal errors or 0-row replacements).\n\n---\n\n## **USER-CENTRIC PROGRESS DISCLOSURE**\nBecause dataset generation and expansion is a long-running operation spanning multiple steps and queries, you must keep the user informed of high-level progress by outputting a progress header as you transition through phases.\n\nYou must prepend this exact block to the very top of every single response you generate.\n\n```text\n### 🧭 Workflow Progress\n* **Milestone:** [Step X of Y: User-Friendly Stage Title]\n* **Status:** [One sentence summarizing what was completed and what is currently running\u002Fnext]\n```\n---\n\n## **INTERNAL PHASES**\n\n### **PHASE 1: ENVIRONMENT & CONTEXT ACQUISITION**\n*   **Goal:** Map the technical and business domain.\n*   **Mandatory Actions:**\n    1.  Read `\u003Cskill_dir>\u002Freferences\u002Fenvironment-context-acquisition.md`.\n    2.  Use MCP tools to list database schemas and identify the `\u003Csource>-execute-sql` tool for validation.\n    3.  Process artifacts to map business concepts to the schema.\n    4.  Establish the output file name (default: `golden.json` if unspecified).\n    5.  Write\u002FUpdate the environment and context acquisition report capturing the domain map, artifact registry, and any business rule shifts detected.\n*   **Exit Criteria:** A `evalset_environment_inputs.md` report is written to disk.\n\n### **PHASE 2: STRATEGIC PLANNING [WAIT FOR USER APPROVAL]**\n*   **Goal:** Create `evalset_gen_plan.md` and get explicit user approval on the dataset requirements.\n*   **Mandatory Actions:**\n    1.  Read `\u003Cskill_dir>\u002Freferences\u002Fgeneration-plan-requirements.md`.\n    2.  **Ensure Robust Dataset Size:** Unless the user has explicitly specified a custom target, the minimum target volume for a NL2SQL dataset is **at least 50 questions**.\n    3.  **Compose and Update Plan (`evalset_gen_plan.md`):** Systematically complete every section required by `generation-plan-requirements.md`. You must write out the plan completely without skipping sections, using placeholders, or abbreviating. Place the main decisions requiring user-review at the top of the plan.\n    4. **[USER APPROVAL GATE]:** STOP. You MUST halt and wait for user approval of `evalset_gen_plan.md`. **DO NOT proceed to the next phase until explicitly given permission.**\n*   **Exit Criteria:** User explicitly approved `evalset_gen_plan.md` and indicated we may proceed to the next phase.\n\n### **PHASE 3: INTELLIGENT GENERATION**\n*   **Goal:** Create the core \"Seed\" dataset with execution-guided proof.\n*   **Mandatory Actions:**\n    1.  Execute workflow in `\u003Cskill_dir>\u002Freferences\u002Fgeneration-cot.md`, saving validated examples via `generate_dataset` MCP Tool to an interim dataset file `temp_golden.json`.\n*   **Exit Criteria:** `temp_golden.json` is created, and every single example in `temp_golden.json` satisfies `evalset_gen_plan.md`'s conditions on the initial seed dataset. \n\n### **PHASE 4: EXPANSION & DIVERSIFICATION**\n*   **Goal:** Increase volume and edge-case coverage to reach the approved target volume.\n*   **Mandatory Actions:**\n    1.  Execute workflow in `\u003Cskill_dir>\u002Freferences\u002Fdataset_expansion.md`, saving validated examples via `generate_dataset` MCP tool to an interim dataset file `temp_golden.json`.\n*   **Exit Criteria:** `temp_golden.json` is updated, and every single example in the expanded dataset satisfies `evalset_gen_plan.md`'s conditions on the expanded dataset. \n\n### **PHASE 5: AUDIT & REPORTING [WAIT FOR USER APPROVAL]**\n*   **Goal:** Assess the quality and diversity of the generated dataset, and get explicit user approval on the dataset.\n*   **Mandatory Actions:**\n    1.  Generate and write audit reports per `\u003Cskill_dir>\u002Freferences\u002Freview-protocol.md`.\n    2.  **[USER APPROVAL GATE]:** STOP. You MUST halt and wait for user approval of the dataset and resolution of all questions before proceeding to the next phase.\n*   **Exit Criteria:** User explicitly approved the dataset and indicated we may proceed to the next phase.\n\n### **PHASE 6: FINALIZATION**\n*   **Goal:** Deliver the final package and any requested subsets to the active working directory.\n*   **Precondition:** All required phase audit reports (environment acquisition, strategic plan, pair-level review, dataset-level review) must exist on disk.\n*   **Mandatory Actions:**\n    1.  **Save Dataset:** Copy the temp dataset file `temp_golden.json` to the `output_file_path` — default to the user's current working directory. If the file already exists, verify whether we should overwrite with the user.\n    2.  **Move Deliverables:** Ensure all written files (`.json`, `.md`, reports) are moved to the user's active directory if they were initially created elsewhere.\n",{"data":36,"body":37},{"name":4,"description":6},{"type":38,"children":39},"root",[40,67,74,81,86,98,107,211,215,224,229,234,246,249,258,268,359,374,499,508,588,597,665,678,734,743],{"type":41,"tag":42,"props":43,"children":44},"element","blockquote",{},[45],{"type":41,"tag":46,"props":47,"children":48},"p",{},[49,65],{"type":41,"tag":50,"props":51,"children":52},"strong",{},[53,56,63],{"type":54,"value":55},"text","Load the ",{"type":41,"tag":57,"props":58,"children":60},"code",{"className":59},[],[61],{"type":54,"value":62},"context-engineering-workflow",{"type":54,"value":64}," skill first.",{"type":54,"value":66}," It holds the shared context this phase depends on: workspace layout, state file conventions, phase order, and safety protocol. Do not proceed with this phase without reading it.",{"type":41,"tag":68,"props":69,"children":71},"h1",{"id":70},"phase-evaluation-dataset-prep-expansion",[72],{"type":54,"value":73},"Phase: Evaluation Dataset Prep & Expansion",{"type":41,"tag":75,"props":76,"children":78},"h2",{"id":77},"goal",[79],{"type":54,"value":80},"Goal",{"type":41,"tag":46,"props":82,"children":83},{},[84],{"type":54,"value":85},"Build a high-quality \"golden\" ground-truth dataset of Natural Language Questions (NLQ) and reference SQL queries for evaluation.",{"type":41,"tag":46,"props":87,"children":88},{},[89,91,96],{"type":54,"value":90},"You are an expert Database Architect, SQL Reverse-Engineering Specialist, and Dataset Evaluation Engineer. Your primary directive is to generate, expand, validate, and sample high-fidelity evaluation datasets (NLQ-SQL pairs) using a ",{"type":41,"tag":50,"props":92,"children":93},{},[94],{"type":54,"value":95},"state-driven, tool-verified, gated workflow",{"type":54,"value":97},".",{"type":41,"tag":75,"props":99,"children":101},{"id":100},"core-operating-principles",[102],{"type":41,"tag":50,"props":103,"children":104},{},[105],{"type":54,"value":106},"CORE OPERATING PRINCIPLES",{"type":41,"tag":108,"props":109,"children":110},"ol",{},[111,145,155,165,175,201],{"type":41,"tag":112,"props":113,"children":114},"li",{},[115,120,122,128,130,136,138,143],{"type":41,"tag":50,"props":116,"children":117},{},[118],{"type":54,"value":119},"Verification",{"type":54,"value":121},": Check for ",{"type":41,"tag":57,"props":123,"children":125},{"className":124},[],[126],{"type":54,"value":127},"tools.yaml",{"type":54,"value":129}," (located in ",{"type":41,"tag":57,"props":131,"children":133},{"className":132},[],[134],{"type":54,"value":135},"autoctx\u002F",{"type":54,"value":137}," for Autoctx workflows) to identify available database configurations. Prompt the user to select the target database for dataset generation. If ",{"type":41,"tag":57,"props":139,"children":141},{"className":140},[],[142],{"type":54,"value":127},{"type":54,"value":144}," is missing, guide the user with init\u002Finit.md to set up.",{"type":41,"tag":112,"props":146,"children":147},{},[148,153],{"type":41,"tag":50,"props":149,"children":150},{},[151],{"type":54,"value":152},"Phase Discipline:",{"type":54,"value":154}," You are strictly forbidden from skipping phases or \"bundling\" multiple phases into a single conversational turn. You must complete the Exit Criteria of one phase before moving to the next.",{"type":41,"tag":112,"props":156,"children":157},{},[158,163],{"type":41,"tag":50,"props":159,"children":160},{},[161],{"type":54,"value":162},"Minimize User Cognitive Load:",{"type":54,"value":164}," For decisions and gates requiring user approval, explicitly specify why the decision matters towards the ultimate goal of curating a high-quality golden dataset. For artifacts requiring user approval, specify what the user should pay closer attention to.",{"type":41,"tag":112,"props":166,"children":167},{},[168,173],{"type":41,"tag":50,"props":169,"children":170},{},[171],{"type":54,"value":172},"Deliverable Persistence & Internal Execution Hiding:",{"type":54,"value":174}," Persist all durable artifacts directly to the file system at the user's working directory rather than outputting text summaries. Hide execution internals from the user: files created for internal quality tracking do not require user awareness or review unless asked.",{"type":41,"tag":112,"props":176,"children":177},{},[178,183,185,191,193,199],{"type":41,"tag":50,"props":179,"children":180},{},[181],{"type":54,"value":182},"Quality & Verification Lock:",{"type":54,"value":184}," Strictly enforce all validation criteria in ",{"type":41,"tag":57,"props":186,"children":188},{"className":187},[],[189],{"type":54,"value":190},"\u003Cskill_dir>\u002Freferences\u002Facceptance-criteria.md",{"type":54,"value":192},", including Zero Hallucination, The Semantic Bridge, and Deterministic SQL ordering (",{"type":41,"tag":57,"props":194,"children":196},{"className":195},[],[197],{"type":54,"value":198},"ORDER BY",{"type":54,"value":200}," tie-breakers).",{"type":41,"tag":112,"props":202,"children":203},{},[204,209],{"type":41,"tag":50,"props":205,"children":206},{},[207],{"type":54,"value":208},"Backtracking:",{"type":54,"value":210}," If a phase reveals quality or correctness issues, you MUST backtrack to a previous phase to fix them (e.g. backtracking to Phase 3 or 4 if Phase 5 audits reveal errors or 0-row replacements).",{"type":41,"tag":212,"props":213,"children":214},"hr",{},[],{"type":41,"tag":75,"props":216,"children":218},{"id":217},"user-centric-progress-disclosure",[219],{"type":41,"tag":50,"props":220,"children":221},{},[222],{"type":54,"value":223},"USER-CENTRIC PROGRESS DISCLOSURE",{"type":41,"tag":46,"props":225,"children":226},{},[227],{"type":54,"value":228},"Because dataset generation and expansion is a long-running operation spanning multiple steps and queries, you must keep the user informed of high-level progress by outputting a progress header as you transition through phases.",{"type":41,"tag":46,"props":230,"children":231},{},[232],{"type":54,"value":233},"You must prepend this exact block to the very top of every single response you generate.",{"type":41,"tag":235,"props":236,"children":241},"pre",{"className":237,"code":239,"language":54,"meta":240},[238],"language-text","### 🧭 Workflow Progress\n* **Milestone:** [Step X of Y: User-Friendly Stage Title]\n* **Status:** [One sentence summarizing what was completed and what is currently running\u002Fnext]\n","",[242],{"type":41,"tag":57,"props":243,"children":244},{"__ignoreMap":240},[245],{"type":54,"value":239},{"type":41,"tag":212,"props":247,"children":248},{},[],{"type":41,"tag":75,"props":250,"children":252},{"id":251},"internal-phases",[253],{"type":41,"tag":50,"props":254,"children":255},{},[256],{"type":54,"value":257},"INTERNAL PHASES",{"type":41,"tag":259,"props":260,"children":262},"h3",{"id":261},"phase-1-environment-context-acquisition",[263],{"type":41,"tag":50,"props":264,"children":265},{},[266],{"type":54,"value":267},"PHASE 1: ENVIRONMENT & CONTEXT ACQUISITION",{"type":41,"tag":269,"props":270,"children":271},"ul",{},[272,282,341],{"type":41,"tag":112,"props":273,"children":274},{},[275,280],{"type":41,"tag":50,"props":276,"children":277},{},[278],{"type":54,"value":279},"Goal:",{"type":54,"value":281}," Map the technical and business domain.",{"type":41,"tag":112,"props":283,"children":284},{},[285,290],{"type":41,"tag":50,"props":286,"children":287},{},[288],{"type":54,"value":289},"Mandatory Actions:",{"type":41,"tag":108,"props":291,"children":292},{},[293,305,318,323,336],{"type":41,"tag":112,"props":294,"children":295},{},[296,298,304],{"type":54,"value":297},"Read ",{"type":41,"tag":57,"props":299,"children":301},{"className":300},[],[302],{"type":54,"value":303},"\u003Cskill_dir>\u002Freferences\u002Fenvironment-context-acquisition.md",{"type":54,"value":97},{"type":41,"tag":112,"props":306,"children":307},{},[308,310,316],{"type":54,"value":309},"Use MCP tools to list database schemas and identify the ",{"type":41,"tag":57,"props":311,"children":313},{"className":312},[],[314],{"type":54,"value":315},"\u003Csource>-execute-sql",{"type":54,"value":317}," tool for validation.",{"type":41,"tag":112,"props":319,"children":320},{},[321],{"type":54,"value":322},"Process artifacts to map business concepts to the schema.",{"type":41,"tag":112,"props":324,"children":325},{},[326,328,334],{"type":54,"value":327},"Establish the output file name (default: ",{"type":41,"tag":57,"props":329,"children":331},{"className":330},[],[332],{"type":54,"value":333},"golden.json",{"type":54,"value":335}," if unspecified).",{"type":41,"tag":112,"props":337,"children":338},{},[339],{"type":54,"value":340},"Write\u002FUpdate the environment and context acquisition report capturing the domain map, artifact registry, and any business rule shifts detected.",{"type":41,"tag":112,"props":342,"children":343},{},[344,349,351,357],{"type":41,"tag":50,"props":345,"children":346},{},[347],{"type":54,"value":348},"Exit Criteria:",{"type":54,"value":350}," A ",{"type":41,"tag":57,"props":352,"children":354},{"className":353},[],[355],{"type":54,"value":356},"evalset_environment_inputs.md",{"type":54,"value":358}," report is written to disk.",{"type":41,"tag":259,"props":360,"children":362},{"id":361},"phase-2-strategic-planning-wait-for-user-approval",[363],{"type":41,"tag":50,"props":364,"children":365},{},[366,368],{"type":54,"value":367},"PHASE 2: STRATEGIC PLANNING ",{"type":41,"tag":369,"props":370,"children":371},"span",{},[372],{"type":54,"value":373},"WAIT FOR USER APPROVAL",{"type":41,"tag":269,"props":375,"children":376},{},[377,394,483],{"type":41,"tag":112,"props":378,"children":379},{},[380,384,386,392],{"type":41,"tag":50,"props":381,"children":382},{},[383],{"type":54,"value":279},{"type":54,"value":385}," Create ",{"type":41,"tag":57,"props":387,"children":389},{"className":388},[],[390],{"type":54,"value":391},"evalset_gen_plan.md",{"type":54,"value":393}," and get explicit user approval on the dataset requirements.",{"type":41,"tag":112,"props":395,"children":396},{},[397,401],{"type":41,"tag":50,"props":398,"children":399},{},[400],{"type":54,"value":289},{"type":41,"tag":108,"props":402,"children":403},{},[404,415,431,456],{"type":41,"tag":112,"props":405,"children":406},{},[407,408,414],{"type":54,"value":297},{"type":41,"tag":57,"props":409,"children":411},{"className":410},[],[412],{"type":54,"value":413},"\u003Cskill_dir>\u002Freferences\u002Fgeneration-plan-requirements.md",{"type":54,"value":97},{"type":41,"tag":112,"props":416,"children":417},{},[418,423,425,430],{"type":41,"tag":50,"props":419,"children":420},{},[421],{"type":54,"value":422},"Ensure Robust Dataset Size:",{"type":54,"value":424}," Unless the user has explicitly specified a custom target, the minimum target volume for a NL2SQL dataset is ",{"type":41,"tag":50,"props":426,"children":427},{},[428],{"type":54,"value":429},"at least 50 questions",{"type":54,"value":97},{"type":41,"tag":112,"props":432,"children":433},{},[434,446,448,454],{"type":41,"tag":50,"props":435,"children":436},{},[437,439,444],{"type":54,"value":438},"Compose and Update Plan (",{"type":41,"tag":57,"props":440,"children":442},{"className":441},[],[443],{"type":54,"value":391},{"type":54,"value":445},"):",{"type":54,"value":447}," Systematically complete every section required by ",{"type":41,"tag":57,"props":449,"children":451},{"className":450},[],[452],{"type":54,"value":453},"generation-plan-requirements.md",{"type":54,"value":455},". You must write out the plan completely without skipping sections, using placeholders, or abbreviating. Place the main decisions requiring user-review at the top of the plan.",{"type":41,"tag":112,"props":457,"children":458},{},[459,469,471,476,478],{"type":41,"tag":50,"props":460,"children":461},{},[462,467],{"type":41,"tag":369,"props":463,"children":464},{},[465],{"type":54,"value":466},"USER APPROVAL GATE",{"type":54,"value":468},":",{"type":54,"value":470}," STOP. You MUST halt and wait for user approval of ",{"type":41,"tag":57,"props":472,"children":474},{"className":473},[],[475],{"type":54,"value":391},{"type":54,"value":477},". ",{"type":41,"tag":50,"props":479,"children":480},{},[481],{"type":54,"value":482},"DO NOT proceed to the next phase until explicitly given permission.",{"type":41,"tag":112,"props":484,"children":485},{},[486,490,492,497],{"type":41,"tag":50,"props":487,"children":488},{},[489],{"type":54,"value":348},{"type":54,"value":491}," User explicitly approved ",{"type":41,"tag":57,"props":493,"children":495},{"className":494},[],[496],{"type":54,"value":391},{"type":54,"value":498}," and indicated we may proceed to the next phase.",{"type":41,"tag":259,"props":500,"children":502},{"id":501},"phase-3-intelligent-generation",[503],{"type":41,"tag":50,"props":504,"children":505},{},[506],{"type":54,"value":507},"PHASE 3: INTELLIGENT GENERATION",{"type":41,"tag":269,"props":509,"children":510},{},[511,520,558],{"type":41,"tag":112,"props":512,"children":513},{},[514,518],{"type":41,"tag":50,"props":515,"children":516},{},[517],{"type":54,"value":279},{"type":54,"value":519}," Create the core \"Seed\" dataset with execution-guided proof.",{"type":41,"tag":112,"props":521,"children":522},{},[523,527],{"type":41,"tag":50,"props":524,"children":525},{},[526],{"type":54,"value":289},{"type":41,"tag":108,"props":528,"children":529},{},[530],{"type":41,"tag":112,"props":531,"children":532},{},[533,535,541,543,549,551,557],{"type":54,"value":534},"Execute workflow in ",{"type":41,"tag":57,"props":536,"children":538},{"className":537},[],[539],{"type":54,"value":540},"\u003Cskill_dir>\u002Freferences\u002Fgeneration-cot.md",{"type":54,"value":542},", saving validated examples via ",{"type":41,"tag":57,"props":544,"children":546},{"className":545},[],[547],{"type":54,"value":548},"generate_dataset",{"type":54,"value":550}," MCP Tool to an interim dataset file ",{"type":41,"tag":57,"props":552,"children":554},{"className":553},[],[555],{"type":54,"value":556},"temp_golden.json",{"type":54,"value":97},{"type":41,"tag":112,"props":559,"children":560},{},[561,565,567,572,574,579,581,586],{"type":41,"tag":50,"props":562,"children":563},{},[564],{"type":54,"value":348},{"type":54,"value":566}," ",{"type":41,"tag":57,"props":568,"children":570},{"className":569},[],[571],{"type":54,"value":556},{"type":54,"value":573}," is created, and every single example in ",{"type":41,"tag":57,"props":575,"children":577},{"className":576},[],[578],{"type":54,"value":556},{"type":54,"value":580}," satisfies ",{"type":41,"tag":57,"props":582,"children":584},{"className":583},[],[585],{"type":54,"value":391},{"type":54,"value":587},"'s conditions on the initial seed dataset.",{"type":41,"tag":259,"props":589,"children":591},{"id":590},"phase-4-expansion-diversification",[592],{"type":41,"tag":50,"props":593,"children":594},{},[595],{"type":54,"value":596},"PHASE 4: EXPANSION & DIVERSIFICATION",{"type":41,"tag":269,"props":598,"children":599},{},[600,609,643],{"type":41,"tag":112,"props":601,"children":602},{},[603,607],{"type":41,"tag":50,"props":604,"children":605},{},[606],{"type":54,"value":279},{"type":54,"value":608}," Increase volume and edge-case coverage to reach the approved target volume.",{"type":41,"tag":112,"props":610,"children":611},{},[612,616],{"type":41,"tag":50,"props":613,"children":614},{},[615],{"type":54,"value":289},{"type":41,"tag":108,"props":617,"children":618},{},[619],{"type":41,"tag":112,"props":620,"children":621},{},[622,623,629,630,635,637,642],{"type":54,"value":534},{"type":41,"tag":57,"props":624,"children":626},{"className":625},[],[627],{"type":54,"value":628},"\u003Cskill_dir>\u002Freferences\u002Fdataset_expansion.md",{"type":54,"value":542},{"type":41,"tag":57,"props":631,"children":633},{"className":632},[],[634],{"type":54,"value":548},{"type":54,"value":636}," MCP tool to an interim dataset file ",{"type":41,"tag":57,"props":638,"children":640},{"className":639},[],[641],{"type":54,"value":556},{"type":54,"value":97},{"type":41,"tag":112,"props":644,"children":645},{},[646,650,651,656,658,663],{"type":41,"tag":50,"props":647,"children":648},{},[649],{"type":54,"value":348},{"type":54,"value":566},{"type":41,"tag":57,"props":652,"children":654},{"className":653},[],[655],{"type":54,"value":556},{"type":54,"value":657}," is updated, and every single example in the expanded dataset satisfies ",{"type":41,"tag":57,"props":659,"children":661},{"className":660},[],[662],{"type":54,"value":391},{"type":54,"value":664},"'s conditions on the expanded dataset.",{"type":41,"tag":259,"props":666,"children":668},{"id":667},"phase-5-audit-reporting-wait-for-user-approval",[669],{"type":41,"tag":50,"props":670,"children":671},{},[672,674],{"type":54,"value":673},"PHASE 5: AUDIT & REPORTING ",{"type":41,"tag":369,"props":675,"children":676},{},[677],{"type":54,"value":373},{"type":41,"tag":269,"props":679,"children":680},{},[681,690,725],{"type":41,"tag":112,"props":682,"children":683},{},[684,688],{"type":41,"tag":50,"props":685,"children":686},{},[687],{"type":54,"value":279},{"type":54,"value":689}," Assess the quality and diversity of the generated dataset, and get explicit user approval on the dataset.",{"type":41,"tag":112,"props":691,"children":692},{},[693,697],{"type":41,"tag":50,"props":694,"children":695},{},[696],{"type":54,"value":289},{"type":41,"tag":108,"props":698,"children":699},{},[700,712],{"type":41,"tag":112,"props":701,"children":702},{},[703,705,711],{"type":54,"value":704},"Generate and write audit reports per ",{"type":41,"tag":57,"props":706,"children":708},{"className":707},[],[709],{"type":54,"value":710},"\u003Cskill_dir>\u002Freferences\u002Freview-protocol.md",{"type":54,"value":97},{"type":41,"tag":112,"props":713,"children":714},{},[715,723],{"type":41,"tag":50,"props":716,"children":717},{},[718,722],{"type":41,"tag":369,"props":719,"children":720},{},[721],{"type":54,"value":466},{"type":54,"value":468},{"type":54,"value":724}," STOP. You MUST halt and wait for user approval of the dataset and resolution of all questions before proceeding to the next phase.",{"type":41,"tag":112,"props":726,"children":727},{},[728,732],{"type":41,"tag":50,"props":729,"children":730},{},[731],{"type":54,"value":348},{"type":54,"value":733}," User explicitly approved the dataset and indicated we may proceed to the next phase.",{"type":41,"tag":259,"props":735,"children":737},{"id":736},"phase-6-finalization",[738],{"type":41,"tag":50,"props":739,"children":740},{},[741],{"type":54,"value":742},"PHASE 6: FINALIZATION",{"type":41,"tag":269,"props":744,"children":745},{},[746,755,765],{"type":41,"tag":112,"props":747,"children":748},{},[749,753],{"type":41,"tag":50,"props":750,"children":751},{},[752],{"type":54,"value":279},{"type":54,"value":754}," Deliver the final package and any requested subsets to the active working directory.",{"type":41,"tag":112,"props":756,"children":757},{},[758,763],{"type":41,"tag":50,"props":759,"children":760},{},[761],{"type":54,"value":762},"Precondition:",{"type":54,"value":764}," All required phase audit reports (environment acquisition, strategic plan, pair-level review, dataset-level review) must exist on disk.",{"type":41,"tag":112,"props":766,"children":767},{},[768,772],{"type":41,"tag":50,"props":769,"children":770},{},[771],{"type":54,"value":289},{"type":41,"tag":108,"props":773,"children":774},{},[775,800],{"type":41,"tag":112,"props":776,"children":777},{},[778,783,785,790,792,798],{"type":41,"tag":50,"props":779,"children":780},{},[781],{"type":54,"value":782},"Save Dataset:",{"type":54,"value":784}," Copy the temp dataset file ",{"type":41,"tag":57,"props":786,"children":788},{"className":787},[],[789],{"type":54,"value":556},{"type":54,"value":791}," to the ",{"type":41,"tag":57,"props":793,"children":795},{"className":794},[],[796],{"type":54,"value":797},"output_file_path",{"type":54,"value":799}," — default to the user's current working directory. If the file already exists, verify whether we should overwrite with the user.",{"type":41,"tag":112,"props":801,"children":802},{},[803,808,810,816,818,824],{"type":41,"tag":50,"props":804,"children":805},{},[806],{"type":54,"value":807},"Move Deliverables:",{"type":54,"value":809}," Ensure all written files (",{"type":41,"tag":57,"props":811,"children":813},{"className":812},[],[814],{"type":54,"value":815},".json",{"type":54,"value":817},", ",{"type":41,"tag":57,"props":819,"children":821},{"className":820},[],[822],{"type":54,"value":823},".md",{"type":54,"value":825},", reports) are moved to the user's active directory if they were initially created elsewhere.",{"items":827,"total":1012},[828,846,860,882,896,907,921,938,955,968,984,994],{"slug":829,"name":829,"fn":830,"description":831,"org":832,"tags":833,"stars":843,"repoUrl":844,"updatedAt":845},"kb-search","search and extract local knowledge base documents","Allows listing, searching and extracting information from local knowledge base documents for information about tables\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[834,837,840],{"name":835,"slug":836,"type":16},"Documentation","documentation",{"name":838,"slug":839,"type":16},"Knowledge Base","knowledge-base",{"name":841,"slug":842,"type":16},"Search","search",6749,"https:\u002F\u002Fgithub.com\u002FGoogleCloudPlatform\u002Fknowledge-catalog","2026-07-12T07:38:52.157375",{"slug":847,"name":848,"fn":849,"description":850,"org":851,"tags":852,"stars":843,"repoUrl":844,"updatedAt":859},"knowledgecatalogdiscoveryagent","knowledge_catalog_discovery_agent","search and rank Knowledge Catalog data entries","Analyzes user queries, extracts relevant predicates, and utilizes Knowledge Catalog Search to find and rank the most relevant data entries. Engages with the user throughout the process.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[853,854,855,858],{"name":18,"slug":19,"type":16},{"name":9,"slug":8,"type":16},{"name":856,"slug":857,"type":16},"Knowledge Management","knowledge-management",{"name":841,"slug":842,"type":16},"2026-07-12T07:38:22.196851",{"slug":861,"name":861,"fn":862,"description":863,"org":864,"tags":865,"stars":879,"repoUrl":880,"updatedAt":881},"contributing","contribute to Cloud Foundation Fabric","End-to-end workflow for contributing to Cloud Foundation Fabric: triaging GitHub issues, proactive feature development, validating with tests and Policy Troubleshooter, and submitting sanitized Pull Requests. Use when addressing a Fabric GitHub issue, developing a module or FAST stage change, or preparing a branch for a pull request.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[866,869,872,875,876],{"name":867,"slug":868,"type":16},"Automation","automation",{"name":870,"slug":871,"type":16},"Engineering","engineering",{"name":873,"slug":874,"type":16},"GitHub","github",{"name":9,"slug":8,"type":16},{"name":877,"slug":878,"type":16},"Pull Requests","pull-requests",2062,"https:\u002F\u002Fgithub.com\u002FGoogleCloudPlatform\u002Fcloud-foundation-fabric","2026-07-31T06:23:36.935005",{"slug":883,"name":883,"fn":884,"description":885,"org":886,"tags":887,"stars":879,"repoUrl":880,"updatedAt":895},"fabric-builder","generate Terraform code for Google Cloud","Generates idiomatic Cloud Foundation Fabric (CFF) Terraform code using CFF modules. Use when users ask to create GCP resources, use Fabric modules, or generate Terraform code for Google Cloud.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[888,889,892],{"name":9,"slug":8,"type":16},{"name":890,"slug":891,"type":16},"Infrastructure as Code","infrastructure-as-code",{"name":893,"slug":894,"type":16},"Terraform","terraform","2026-07-12T07:38:23.514555",{"slug":897,"name":897,"fn":898,"description":899,"org":900,"tags":901,"stars":879,"repoUrl":880,"updatedAt":906},"fast-0-org-setup-prereqs","prepare prerequisites for FAST 0-org-setup","Guides the user step-by-step through the prerequisites for the FAST 0-org-setup stage, supporting both Standard GCP and Google Cloud Dedicated (GCD) environments. Use when a user asks to prepare or run prerequisites for 0-org-setup or bootstrap the FAST landing zone.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[902,903],{"name":9,"slug":8,"type":16},{"name":904,"slug":905,"type":16},"Operations","operations","2026-07-12T07:38:28.127148",{"slug":908,"name":908,"fn":909,"description":910,"org":911,"tags":912,"stars":918,"repoUrl":919,"updatedAt":920},"agent-aware-cli","design agent-aware command-line interfaces","Guide for designing and implementing command-line interfaces (CLIs) that are equally usable by human developers and automated coding agents. Use when the user wants to build a CLI, apply CLI best practices, or use Go with Cobra and Viper.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[913,916,917],{"name":914,"slug":915,"type":16},"CLI","cli",{"name":870,"slug":871,"type":16},{"name":9,"slug":8,"type":16},1150,"https:\u002F\u002Fgithub.com\u002FGoogleCloudPlatform\u002Fvertex-ai-creative-studio","2026-07-12T07:39:08.41406",{"slug":922,"name":922,"fn":923,"description":924,"org":925,"tags":926,"stars":918,"repoUrl":919,"updatedAt":937},"build-mcp-genmedia","build and configure GenAI MCP servers","Builds the mcp-genmedia Go MCP servers (nanobanana, veo, lyria, gemini-multimodal, chirp3-hd, avtool) from source and wires them into settings.json. Use this skill whenever the MCP tools are missing or broken — typically at the start of a new session, after a container restart, or when \u002Ftmp has been wiped. The prebuilt binaries in \u002Fworkspace\u002F.local\u002Fbin\u002F have no exec bit and live on a noexec mount; this skill compiles fresh executables into \u002Ftmp\u002Fbin\u002F where execution is allowed.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[927,930,931,934],{"name":928,"slug":929,"type":16},"API Development","api-development",{"name":9,"slug":8,"type":16},{"name":932,"slug":933,"type":16},"LLM","llm",{"name":935,"slug":936,"type":16},"MCP","mcp","2026-07-12T07:39:10.911302",{"slug":939,"name":939,"fn":940,"description":941,"org":942,"tags":943,"stars":918,"repoUrl":919,"updatedAt":954},"genmedia-audio-engineer","synthesize and mix audio content","Expert in audio synthesis, music generation, and mixing. Use when creating podcasts, background scores, or multi-track audio layering using mcp-chirp3-go, mcp-lyria-go, mcp-gemini-go, mcp-nanobanana-go, and mcp-avtool-go.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[944,947,950,951],{"name":945,"slug":946,"type":16},"Audio","audio",{"name":948,"slug":949,"type":16},"Creative","creative",{"name":9,"slug":8,"type":16},{"name":952,"slug":953,"type":16},"Vertex AI","vertex-ai","2026-07-12T07:39:16.623879",{"slug":956,"name":956,"fn":957,"description":958,"org":959,"tags":960,"stars":918,"repoUrl":919,"updatedAt":967},"genmedia-image-artist","generate and edit AI images","Expert in AI image generation and editing. Use when the user needs high-quality textures, character-consistent visuals, or image-to-image editing using mcp-nanobanana-go.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[961,962,963,966],{"name":948,"slug":949,"type":16},{"name":9,"slug":8,"type":16},{"name":964,"slug":965,"type":16},"Image Generation","image-generation",{"name":952,"slug":953,"type":16},"2026-07-12T07:39:15.372822",{"slug":969,"name":969,"fn":970,"description":971,"org":972,"tags":973,"stars":918,"repoUrl":919,"updatedAt":983},"genmedia-producer","produce multi-step media content","Expert media production assistant. Use when requested to help with storyboarding, podcast creation, audio assembly, or complex multi-step media workflows using the GenMedia MCP servers (Veo, Lyria, Gemini TTS, NanoBanana).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[974,975,976,977,980],{"name":945,"slug":946,"type":16},{"name":948,"slug":949,"type":16},{"name":9,"slug":8,"type":16},{"name":978,"slug":979,"type":16},"Media","media",{"name":981,"slug":982,"type":16},"Video","video","2026-07-12T07:39:09.672849",{"slug":985,"name":985,"fn":986,"description":987,"org":988,"tags":989,"stars":918,"repoUrl":919,"updatedAt":993},"genmedia-video-editor","edit and compose video content","Expert in video composition, editing, and format conversion. Use when the user wants to generate high-quality video, overlay images on video, concatenate clips, create GIFs, or sync audio to video using mcp-avtool-go and mcp-veo-go.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[990,991,992],{"name":948,"slug":949,"type":16},{"name":9,"slug":8,"type":16},{"name":981,"slug":982,"type":16},"2026-07-12T07:39:13.749081",{"slug":995,"name":995,"fn":996,"description":997,"org":998,"tags":999,"stars":918,"repoUrl":919,"updatedAt":1011},"genmedia-voice-director","generate expressive text-to-speech with Gemini","Expert in casting, directing, and generating expressive text-to-speech using Gemini TTS. Use this when the user needs virtual voice actor personas, expressive speech generation, or multiple variations of a voiceover (like \"take 3 on the bounce\").",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1000,1001,1002,1005,1008],{"name":945,"slug":946,"type":16},{"name":948,"slug":949,"type":16},{"name":1003,"slug":1004,"type":16},"Gemini","gemini",{"name":1006,"slug":1007,"type":16},"Speech","speech",{"name":1009,"slug":1010,"type":16},"Text-to-Speech","text-to-speech","2026-07-12T07:39:17.86673",80,{"items":1014,"total":1099},[1015,1032,1039,1050,1063,1076,1088],{"slug":1016,"name":1016,"fn":1017,"description":1018,"org":1019,"tags":1020,"stars":24,"repoUrl":25,"updatedAt":1031},"context-engineering-bootstrap","bootstrap initial ContextSet for database enrichment","Guides the agent to bootstrap an initial ContextSet (templates, facets, and value searches) by deducing key information from the database schema and generating a ContextSet file.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1021,1024,1027,1030],{"name":1022,"slug":1023,"type":16},"Context","context",{"name":1025,"slug":1026,"type":16},"Data Modeling","data-modeling",{"name":1028,"slug":1029,"type":16},"Database","database",{"name":9,"slug":8,"type":16},"2026-07-18T05:47:56.797107",{"slug":4,"name":4,"fn":5,"description":6,"org":1033,"tags":1034,"stars":24,"repoUrl":25,"updatedAt":26},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1035,1036,1037,1038],{"name":18,"slug":19,"type":16},{"name":14,"slug":15,"type":16},{"name":9,"slug":8,"type":16},{"name":21,"slug":22,"type":16},{"slug":1040,"name":1040,"fn":1041,"description":1042,"org":1043,"tags":1044,"stars":24,"repoUrl":25,"updatedAt":1049},"context-engineering-evaluate","evaluate ContextSet against golden datasets","Guides the agent to execute an evaluation of a ContextSet against a golden NLQ+SQL dataset using the Evalbench framework.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1045,1046,1047,1048],{"name":18,"slug":19,"type":16},{"name":14,"slug":15,"type":16},{"name":9,"slug":8,"type":16},{"name":21,"slug":22,"type":16},"2026-07-18T05:47:52.594607",{"slug":1051,"name":1051,"fn":1052,"description":1053,"org":1054,"tags":1055,"stars":24,"repoUrl":25,"updatedAt":1062},"context-engineering-hillclimb","improve ContextSet via hill-climbing iterations","Guides the agent to perform hill-climbing iterations to improve a ContextSet based on Evalbench evaluation results.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1056,1057,1058,1059],{"name":18,"slug":19,"type":16},{"name":14,"slug":15,"type":16},{"name":9,"slug":8,"type":16},{"name":1060,"slug":1061,"type":16},"Optimization","optimization","2026-07-18T05:47:54.951176",{"slug":1064,"name":1064,"fn":1065,"description":1066,"org":1067,"tags":1068,"stars":24,"repoUrl":25,"updatedAt":1075},"context-engineering-init","initialize context engineering and database connections","Orchestrates the initialization workflow for context engineering, and provides helper workflow for setting up database connections by creating or updating tools.yaml configurations.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1069,1072,1073,1074],{"name":1070,"slug":1071,"type":16},"Configuration","configuration",{"name":1028,"slug":1029,"type":16},{"name":870,"slug":871,"type":16},{"name":9,"slug":8,"type":16},"2026-07-18T05:47:59.493278",{"slug":62,"name":62,"fn":1077,"description":1078,"org":1079,"tags":1080,"stars":24,"repoUrl":25,"updatedAt":1087},"optimize context sets for Gemini data agents","Context engineering for Gemini Data Analytics API's data agent developer platform tools. Generates, evaluates, and iteratively optimizes a ContextSet (Templates, Facets, Value Searches) to maximize Natural-Language-to-SQL translation accuracy. Use this skill to run the automated setup, NL-SQL pair evaluation dataset generation and expansion, bootstrapping, scoring, and optimization pipeline. For manual authoring standards and schema syntax rules, see the context-generation-guide skill.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1081,1084,1085,1086],{"name":1082,"slug":1083,"type":16},"Agent Context","agent-context",{"name":18,"slug":19,"type":16},{"name":1003,"slug":1004,"type":16},{"name":9,"slug":8,"type":16},"2026-07-18T05:13:44.430929",{"slug":1089,"name":1089,"fn":1090,"description":1091,"org":1092,"tags":1093,"stars":24,"repoUrl":25,"updatedAt":1098},"context-generation-guide","generate context items for database enrichment","Guidelines and best practices for generating context items (Templates, Facets, Value Searches). Use this skill whenever the user asks to create, author, or generate context for database enrichment, or asks for examples and instructions on how to write templates, facets, or value searches. It helps bridge the gap between LLMs and structured databases. For running the automated generation, evaluation, and tuning lifecycle, see the context-engineering-workflow skill.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1094,1095,1096,1097],{"name":1082,"slug":1083,"type":16},{"name":1028,"slug":1029,"type":16},{"name":835,"slug":836,"type":16},{"name":9,"slug":8,"type":16},"2026-07-18T05:13:43.403052",7]