[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M75NADB1M4Q2XAEAN0M2FQWS":3,"skill-same:01M75NADB1M4Q2XAEAN0M2FQWS":44},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":25,"commitSha":26,"skillDirectory":27,"forksCount":14,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":43},"01M75NADB1M4Q2XAEAN0M2FQWS","prd",null,"Use this to turn a rough feature idea into a reviewable PRD inside a Kata run. Grills the operator one focused question at a time — problem, users, scope, success criteria, non-goals — through the `ask_user` tool, challenges vague answers, and writes the finished PRD to a file. Requirements only; no design, no code.","satish-krishna\u002Fkata","satish-krishna","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F26695756?v=4","01M75N910J6TF3ZF2K71GYGEV6","SatishKrishna Pilla","01M75N91142BHTH4G6E6ZXCM9F",0,"MIT","https:\u002F\u002Fgithub.com\u002Fsatish-krishna\u002Fkata\u002Fblob\u002Fmain\u002F.claude\u002Fplugins\u002Fkata\u002Fskills\u002Fprd\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: prd\ndescription: Use this to turn a rough feature idea into a reviewable PRD inside a Kata run. Grills the operator one focused question at a time — problem, users, scope, success criteria, non-goals — through the `ask_user` tool, challenges vague answers, and writes the finished PRD to a file. Requirements only; no design, no code.\nuser-invocable: true\n---\n\n# PRD (grill me)\n\nInterrogate a rough idea until it is a product requirements document someone could plan against. You are the skeptical interviewer: your job is to extract what the operator actually needs, not to transcribe what they first said. Requirements are the deliverable — no architecture, no plan, no code.\n\n## The one rule that changes everything\n\n**`ask_user` is your only mouth. A question you write as prose never reaches the operator.** A Kata run drives `claude -p` headlessly; end a turn with a plain-text question and the run simply ends, unanswered. Everywhere this skill says \"ask\", call the `ask_user` tool — it pauses the run, shows the question, and returns the answer in the same turn. Never use the built-in `AskUserQuestion`; it is disabled in Kata runs.\n\nAsk one focused question per call (batch only genuinely-related ones). Prefer `select`\u002F`confirm` with your recommended option first over open `text` whenever the choices are knowable. Read each answer before forming the next question. Kata wires `ask_user` into every run, interactive or not, so ask whenever you need the operator. Only if the tool is genuinely missing from your toolset, do not stall: make the conservative assumption, record it in the PRD under **Assumptions**, and continue.\n\n## Model policy\n\nYou (the driver) run on Sonnet and do the judgment work — questioning, challenging, synthesizing. Dispatch mechanical reading to the plugin's **`kata-scout`** subagent (pinned to Haiku): repo survey, finding prior art, extracting conventions. Do not burn driver turns on file-tree tours.\n\n## The flow\n\nCreate a task per step and work them in order:\n\n1. **Recon before questions.** Dispatch `kata-scout` to survey what exists: the product's shape, any related features, docs, prior PRDs and their format. Never spend a question on something the repo already answers.\n2. **Grill.** One `ask_user` question at a time, roughly in this ladder — skip rungs the answers have already settled:\n   - the problem: who has it, how often, what it costs them today;\n   - the current workaround, and why it is not good enough;\n   - success: what observably changes when this ships, and how it's measured;\n   - users and their jobs: who touches this, to do what;\n   - scope: what is in, and — pressed explicitly — what is *out*;\n   - constraints: compatibility, performance, deadlines, platforms;\n   - risks and edge cases the operator is worried about.\n\n   Grill means grill: when an answer is vague (\"it should be fast\"), ask for the number. When two answers conflict, surface the conflict and make them pick. When everything is \"must have\", force a must \u002F should \u002F won't split. Stop only when you could defend every requirement to a stranger.\n3. **Draft.** Assemble the PRD: Problem; Goals & success metrics; Users & jobs; Requirements (must \u002F should \u002F won't, each testable); Non-goals; Constraints & assumptions; Risks & open questions; Acceptance criteria.\n4. **Confirm.** Present a tight summary through `ask_user` (`select`: approve \u002F amend, approve first). On amend, loop back to the gap.\n5. **Deliver — to a file.** Resolve the destination in order: a path the task names wins; otherwise ask where to write through `ask_user` — a `select` leading with the default `docs\u002Fprds\u002F\u003CYYYY-MM-DD>-\u003Ctopic>-prd.md`, then any doc convention the recon surfaced, then an \"elsewhere\" option whose answer is a custom path. A run's final message is not a durable artifact — the file is. Summarize in your closing message. Do not commit unless the task says to.\n\n## Anti-patterns\n\n- **Asking in prose and stopping.** The run ends; the operator sees nothing.\n- **Transcribing instead of grilling.** \"Users want it faster\" is not a requirement. Push for numbers, priorities, and exclusions.\n- **Dumping a questionnaire.** One focused question per call; this is a dialogue, not a form.\n- **Sneaking in design.** \"Add a `--flag` to the CLI\" is a solution. Capture the need; leave the how to `kata:plan`.\n- **Delivering only in the closing message.** The file is the deliverable.\n","# PRD (grill me)\n\nInterrogate a rough idea until it is a product requirements document someone could plan against. You are the skeptical interviewer: your job is to extract what the operator actually needs, not to transcribe what they first said. Requirements are the deliverable — no architecture, no plan, no code.\n\n## The one rule that changes everything\n\n**`ask_user` is your only mouth. A question you write as prose never reaches the operator.** A Kata run drives `claude -p` headlessly; end a turn with a plain-text question and the run simply ends, unanswered. Everywhere this skill says \"ask\", call the `ask_user` tool — it pauses the run, shows the question, and returns the answer in the same turn. Never use the built-in `AskUserQuestion`; it is disabled in Kata runs.\n\nAsk one focused question per call (batch only genuinely-related ones). Prefer `select`\u002F`confirm` with your recommended option first over open `text` whenever the choices are knowable. Read each answer before forming the next question. Kata wires `ask_user` into every run, interactive or not, so ask whenever you need the operator. Only if the tool is genuinely missing from your toolset, do not stall: make the conservative assumption, record it in the PRD under **Assumptions**, and continue.\n\n## Model policy\n\nYou (the driver) run on Sonnet and do the judgment work — questioning, challenging, synthesizing. Dispatch mechanical reading to the plugin's **`kata-scout`** subagent (pinned to Haiku): repo survey, finding prior art, extracting conventions. Do not burn driver turns on file-tree tours.\n\n## The flow\n\nCreate a task per step and work them in order:\n\n1. **Recon before questions.** Dispatch `kata-scout` to survey what exists: the product's shape, any related features, docs, prior PRDs and their format. Never spend a question on something the repo already answers.\n2. **Grill.** One `ask_user` question at a time, roughly in this ladder — skip rungs the answers have already settled:\n   - the problem: who has it, how often, what it costs them today;\n   - the current workaround, and why it is not good enough;\n   - success: what observably changes when this ships, and how it's measured;\n   - users and their jobs: who touches this, to do what;\n   - scope: what is in, and — pressed explicitly — what is *out*;\n   - constraints: compatibility, performance, deadlines, platforms;\n   - risks and edge cases the operator is worried about.\n\n   Grill means grill: when an answer is vague (\"it should be fast\"), ask for the number. When two answers conflict, surface the conflict and make them pick. When everything is \"must have\", force a must \u002F should \u002F won't split. Stop only when you could defend every requirement to a stranger.\n3. **Draft.** Assemble the PRD: Problem; Goals & success metrics; Users & jobs; Requirements (must \u002F should \u002F won't, each testable); Non-goals; Constraints & assumptions; Risks & open questions; Acceptance criteria.\n4. **Confirm.** Present a tight summary through `ask_user` (`select`: approve \u002F amend, approve first). On amend, loop back to the gap.\n5. **Deliver — to a file.** Resolve the destination in order: a path the task names wins; otherwise ask where to write through `ask_user` — a `select` leading with the default `docs\u002Fprds\u002F\u003CYYYY-MM-DD>-\u003Ctopic>-prd.md`, then any doc convention the recon surfaced, then an \"elsewhere\" option whose answer is a custom path. A run's final message is not a durable artifact — the file is. Summarize in your closing message. Do not commit unless the task says to.\n\n## Anti-patterns\n\n- **Asking in prose and stopping.** The run ends; the operator sees nothing.\n- **Transcribing instead of grilling.** \"Users want it faster\" is not a requirement. Push for numbers, priorities, and exclusions.\n- **Dumping a questionnaire.** One focused question per call; this is a dialogue, not a form.\n- **Sneaking in design.** \"Add a `--flag` to the CLI\" is a solution. Capture the need; leave the how to `kata:plan`.\n- **Delivering only in the closing message.** The file is the deliverable.\n",{"name":5,"description":7,"user-invocable":24},true,"main","4da69517ed5e3a33f87d8110ee59cff71256ad1f",".claude\u002Fplugins\u002Fkata\u002Fskills\u002Fprd","https:\u002F\u002Fgithub.com\u002Fsatish-krishna\u002Fkata","https:\u002F\u002Fgithub.com\u002Fsatish-krishna","2026-11-14 03:41:14",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":24},[33,36,39],{"tool":34,"command":35},"npx","npx skills add satish-krishna\u002Fkata --skill prd",{"tool":37,"command":38},"bunx","bunx skills add satish-krishna\u002Fkata --skill prd",{"tool":40,"command":41},"pnpm","pnpm dlx skills add satish-krishna\u002Fkata --skill prd","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fsatish-krishna\u002Fkata\u002Fblob\u002Fmain\u002F.claude\u002Fplugins\u002Fkata\u002Fskills\u002Fprd\u002FSKILL.md",[],[45,52,59,66,73,78],{"skillNo":46,"name":47,"nameZh":6,"description":48,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":49,"categories":50,"keywords":51,"publishedAt":19},"01M75NC157R07ZE86MTXWMXM96","probe","Use when the task says to invoke the probe skill. Test fixture only.","https:\u002F\u002Fgithub.com\u002Fsatish-krishna\u002Fkata\u002Fblob\u002Fmain\u002Fcrates\u002Fkata-core\u002Ftests\u002Ffixtures\u002Fhook-probe\u002Fskills\u002Fprobe\u002FSKILL.md",[],[],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":19},"01M75NBZBBE2YNKSZV8YVDH8NB","proof-skill","Proof skill for the web folder. Use whenever editing files under web\u002F. Marker PROOF-SKILL-NESTED-7431.","https:\u002F\u002Fgithub.com\u002Fsatish-krishna\u002Fkata\u002Fblob\u002Fmain\u002Fcrates\u002Fkata-core\u002Ftests\u002Ffixtures\u002Fbuild-proof\u002Fapp\u002Fweb\u002F.claude\u002Fskills\u002Fproof-skill\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":19},"01M75NBTAVF50RK8YJWDG1F048","spartan","Manages spartan\u002Fui, the Angular UI library - adding, composing, fixing, debugging, and styling UI with the Brain (headless primitives) and Helm (styled) layers. Provides project context, component APIs, and usage examples. Applies when working with spartan\u002Fui, @spartan-ng\u002Fbrain, @spartan-ng\u002Fhelm, the @spartan-ng\u002Fcli generators, or any Angular project with a components.json file. Also triggers for \"spartan init\", \"add a spartan component\", or \"set up spartan\u002Fui\".","https:\u002F\u002Fgithub.com\u002Fsatish-krishna\u002Fkata\u002Fblob\u002Fmain\u002Fapp\u002F.claude\u002Fskills\u002Fspartan\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M75NBP134GH4PPK9SAMXN9Y9","house-style","The house frontend conventions for this repo, layered on top of the SpartanNG docs. Read before writing or editing any template or component stylesheet. Covers the layout grammar (grid for regions, flex for inline runs, no nested flex faking a grid) and the no-raw-literals rule for hand-written CSS. These extend the spartan docs where spartan is silent; they never contradict them.","https:\u002F\u002Fgithub.com\u002Fsatish-krishna\u002Fkata\u002Fblob\u002Fmain\u002Fapp\u002F.claude\u002Fskills\u002Fhouse-style\u002FSKILL.md",[],[],{"skillNo":74,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M75NBHCP6EV6KTDR7Y2PFS2R","https:\u002F\u002Fgithub.com\u002Fsatish-krishna\u002Fkata\u002Fblob\u002Fmain\u002Fapp\u002F.agents\u002Fskills\u002Fspartan\u002FSKILL.md",[],[],{"skillNo":79,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M75NBC9BTZKQCFKC0A3PWB7E","https:\u002F\u002Fgithub.com\u002Fsatish-krishna\u002Fkata\u002Fblob\u002Fmain\u002Fapp\u002F.agents\u002Fskills\u002Fhouse-style\u002FSKILL.md",[],[]]