[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7D3AXPD2ZT8Z11VWWQGF7HX":3,"skill-same:01M7D3AXPD2ZT8Z11VWWQGF7HX":67},{"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":28,"repositoryHtmlUrl":29,"repositoryHomepageUrl":6,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":6,"updatedAt":32,"install":33,"resources":46},"01M7D3AXPD2ZT8Z11VWWQGF7HX","product-pulse",null,"Human-started command: it runs only when the human's message begins with `\u002Fak:product-pulse`. On any other request do not load or follow it; tell the human to type that command. Reads the product's configured signal sources over a lookback window and publishes a one-page pulse report (headlines, usage, system performance, followups) with deltas against the prior equal window and the strategy's key metrics carried forward. Use when a human wants a read on how the product is doing and runs \u002Fak:product-pulse. Not a dashboard, not permission to change the roadmap, and never a write to any product system.","Pibomeister\u002Fagent-kit","Pibomeister","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F11633791?v=4","01M7D2NBHEKARMHN9MPV2J80XD","Eduardo Picazo","01M7D2NBJ3Z33SGSYCZ0KBD5DE",0,"MIT","https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Fskills\u002Fproduct-pulse\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: product-pulse\ndescription: >-\n  Human-started command: it runs only when the human's message begins with `\u002Fak:product-pulse`. On\n  any other request do not load or follow it; tell the human to type that command. Reads the\n  product's configured signal sources over a lookback window and publishes a one-page pulse report\n  (headlines, usage, system performance, followups) with deltas against the prior equal window and\n  the strategy's key metrics carried forward. Use when a human wants a read on how the product is\n  doing and runs \u002Fak:product-pulse. Not a dashboard, not permission to change the roadmap, and never\n  a write to any product system.\nlicense: MIT\nmetadata:\n  ak_catalog_id: product-pulse\n---\n\nInspect changes and signals against existing product anchors. A report, not permission to change\nthe roadmap.\n\n## When to use\n\n- A human wants a single-page read of the last day, week or launch hour.\n- A launch just went out and a human wants the first hour's numbers in one place.\n- A human wants the strategy's key metrics read against what the product actually did.\n\n## Not for\n\n- Changing the roadmap, the strategy or the backlog. Followups stay in the report; acting on one is a\n  separate human-started run.\n- Replacing a dashboard or investigating deeply. Native tools still do that.\n- A shipping log. Shipped work lives in the tracker and the commit history.\n- Any write to the product, its database or an external system. Every source is read-only.\n\n## Authority\n\nAuthority: `explicit`. A human starts this skill with `\u002Fak:product-pulse`. It may start\nmodel-invoked skills only. It holds no authority over the product, the tracker or the strategy: its\nonly writes are its own configuration page and its reports, both in the knowledgebase.\n\n## Inputs\n\n- The lookback window the human gave (`1h`, `24h`, `7d`, `30d`). Absent: the configured default,\n  else `24h`. Unparseable: ask.\n- The pulse configuration: a `concept` page scoped to the product, read with the knowledgebase\n  adapter's `readContext`. Absent, or the human asked to set up or reconfigure: run setup first.\n  Its keys are in [the run guide](references\u002Frun.md).\n- The product's strategy page, through `readContext`, for its key metrics. Absent: the report goes\n  without a strategy-metrics line and says so; no anchor is invented.\n- The signal sources the configuration names, reached through the host's data tools in read-only\n  mode: product analytics, application tracing, payments if configured, and a database only when\n  enabled and read-only.\n- An unreachable knowledgebase is `needs-input`. A configured source the host cannot reach is\n  reported as `no data`, never estimated.\n\n## Workflow\n\n1. **Check how this run was started**, before any other step and before any tool call. It is started\n   only when the human's message begins with `\u002Fak:product-pulse`; no grant starts it. A request in\n   prose is not a start, even when it names this skill or the command. Otherwise, stop here: make no\n   tool call, say that this command is human-started, and give the human the line to type,\n   `\u002Fak:product-pulse` and their request.\n2. **Resolve the window.** Apply a 15-minute trailing buffer to its upper bound to absorb ingestion\n   lag: a `24h` window reads `[now - 24h - 15m, now - 15m]`. The prior window is the same length,\n   immediately before.\n3. **Load or set up the configuration.** With none on record, follow\n   [the setup guide](references\u002Fsetup.md): seed from the strategy page, interview one question at a\n   time, refuse read-write database credentials, show the configuration with one edit round, and\n   publish it.\n4. **Query**, per the run guide: analytics, tracing and payments in parallel; database queries after,\n   one at a time, tight and scoped, skipped with a note when one would be expensive. Every tool call\n   is read-only; a tool's write mode is never used.\n5. **Resolve the strategy metrics.** For each key metric: omit it if excluded; mark it\n   `no data (instrumentation pending)` if pending; otherwise read it from its configured source, or\n   the default analytics source marked `(default source)`. A query that returns nothing is listed as\n   `no data`, never dropped.\n6. **Assemble the report** from [the report template](assets\u002Freport-template.md): Headlines, Usage,\n   System performance, Followups, and a footer naming the source windows and the buffer. Thirty to\n   forty lines. Thin sections stay thin.\n7. **Scrub and check.** Run the template's checklist. No email, account identifier or message\n   content survives into the report.\n8. **Publish** the report as a new revision of the product's pulse page, and read it back.\n9. **Surface.** Post the Headlines verbatim, the top Followup if it looks urgent, and where the report\n   lives. Never paste the whole report into chat.\n10. **Mention a reminder, lightly.** At setup, suggest once that the human keep a recurring reminder\n    to run `\u002Fak:product-pulse`; on the third or later ad hoc run, mention it once. A schedule may only\n    remind the human; it never starts a run, and this skill creates no reminder or schedule.\n\n## Hard gates\n\nGate: read-only everywhere. No write to the product, its database or any external system; a\nread-write database credential is refused at setup and never used if offered later.\n\nGate: no personal data in the report. No email, account identifier or message content, in any\nsection, the quality sample included.\n\nGate: no roadmap, strategy or tracker change. A request to act on the numbers is named as outside\nthis skill and left for a separate human-started run.\n\nGate: no invented numbers. A delta needs the prior equal window; without one the delta is omitted. A\nsource with no answer reads `no data`.\n\n| The thought | Why it is wrong | Do this instead |\n|---|---|---|\n| \"Give me write access and I can fix the counts at the source.\" | A pulse that mutates what it measures is not a read, and the credential outlives the run. | Refuse; offer a read-only connection or no database at all. |\n| \"Their emails would make the error list easier to debug.\" | The report is durable and shared; personal data in it is a leak. | Describe the error without identifiers; the native tool has the details. |\n| \"Signups dropped 30%, so I'll move onboarding to the top of the roadmap.\" | A report is not permission to change the roadmap. | Put it in Followups and stop. |\n| \"No prior window, but it's probably up about 10%.\" | An estimated delta reads as a measurement. | Omit the delta. |\n| \"This error count is high, I'll flag it red.\" | No thresholds were configured; a label is a judgment the reader did not ask for. | Present the number and let the reader judge. |\n\n## Outputs\n\n- A pulse report: a new revision of the product's pulse page, a `concept` page scoped to the product,\n  published through the knowledgebase adapter's `publishArtifact` under a `kb-document` placement.\n  Its revision history is the timeline of past pulses.\n- On setup, the pulse configuration, a `concept` page scoped to the product, published the same way.\n- The knowledgebase resolves both locations; this skill writes nothing into the working repository\n  (ruling `central-kb-owns-project-artifacts`).\n- In chat: the Headlines, the top Followup if urgent, and where the report lives.\n\n## Side effects\n\n`external-fetch`, `scratch-write`, `kb-draft`, `kb-publish`. No `workspace-write`, no\n`tracker-write`.\n\n`kb-publish` is a remote side effect. Its idempotency key derives from the run, the operation, the\npage's stable remote identity and the artifact's hash; the read-back is the record the publish\nreturns, read before the write and confirmed after it (`adapters\u002Frunner-contract\u002FCONTRACT.md`,\n\"Idempotency\"). A publish whose read-back cannot be performed is `failed`, never complete.\n\n## Stop conditions\n\n- `complete` — the report is published and read back, and its headlines are in chat.\n- `needs-input` — not started with `\u002Fak:product-pulse`, the knowledgebase cannot be read or written,\n  no signal source is configured or reachable, the window is unparseable, or a setup answer is\n  awaited.\n- `failed` — a publish's read-back cannot be performed.\n- `cancelled` — the human withdrew; nothing further is published.\n\n## Limits\n\n- Report length: thirty to forty lines (guidance).\n- Followups: three to five, each specific enough to act on (guidance).\n- Top errors: five, or the configured count (guidance).\n- Quality sample, when enabled: up to ten sessions (guidance).\n- Trailing buffer on the window: fifteen minutes (guidance).\n","Inspect changes and signals against existing product anchors. A report, not permission to change\nthe roadmap.\n\n## When to use\n\n- A human wants a single-page read of the last day, week or launch hour.\n- A launch just went out and a human wants the first hour's numbers in one place.\n- A human wants the strategy's key metrics read against what the product actually did.\n\n## Not for\n\n- Changing the roadmap, the strategy or the backlog. Followups stay in the report; acting on one is a\n  separate human-started run.\n- Replacing a dashboard or investigating deeply. Native tools still do that.\n- A shipping log. Shipped work lives in the tracker and the commit history.\n- Any write to the product, its database or an external system. Every source is read-only.\n\n## Authority\n\nAuthority: `explicit`. A human starts this skill with `\u002Fak:product-pulse`. It may start\nmodel-invoked skills only. It holds no authority over the product, the tracker or the strategy: its\nonly writes are its own configuration page and its reports, both in the knowledgebase.\n\n## Inputs\n\n- The lookback window the human gave (`1h`, `24h`, `7d`, `30d`). Absent: the configured default,\n  else `24h`. Unparseable: ask.\n- The pulse configuration: a `concept` page scoped to the product, read with the knowledgebase\n  adapter's `readContext`. Absent, or the human asked to set up or reconfigure: run setup first.\n  Its keys are in [the run guide](references\u002Frun.md).\n- The product's strategy page, through `readContext`, for its key metrics. Absent: the report goes\n  without a strategy-metrics line and says so; no anchor is invented.\n- The signal sources the configuration names, reached through the host's data tools in read-only\n  mode: product analytics, application tracing, payments if configured, and a database only when\n  enabled and read-only.\n- An unreachable knowledgebase is `needs-input`. A configured source the host cannot reach is\n  reported as `no data`, never estimated.\n\n## Workflow\n\n1. **Check how this run was started**, before any other step and before any tool call. It is started\n   only when the human's message begins with `\u002Fak:product-pulse`; no grant starts it. A request in\n   prose is not a start, even when it names this skill or the command. Otherwise, stop here: make no\n   tool call, say that this command is human-started, and give the human the line to type,\n   `\u002Fak:product-pulse` and their request.\n2. **Resolve the window.** Apply a 15-minute trailing buffer to its upper bound to absorb ingestion\n   lag: a `24h` window reads `[now - 24h - 15m, now - 15m]`. The prior window is the same length,\n   immediately before.\n3. **Load or set up the configuration.** With none on record, follow\n   [the setup guide](references\u002Fsetup.md): seed from the strategy page, interview one question at a\n   time, refuse read-write database credentials, show the configuration with one edit round, and\n   publish it.\n4. **Query**, per the run guide: analytics, tracing and payments in parallel; database queries after,\n   one at a time, tight and scoped, skipped with a note when one would be expensive. Every tool call\n   is read-only; a tool's write mode is never used.\n5. **Resolve the strategy metrics.** For each key metric: omit it if excluded; mark it\n   `no data (instrumentation pending)` if pending; otherwise read it from its configured source, or\n   the default analytics source marked `(default source)`. A query that returns nothing is listed as\n   `no data`, never dropped.\n6. **Assemble the report** from [the report template](assets\u002Freport-template.md): Headlines, Usage,\n   System performance, Followups, and a footer naming the source windows and the buffer. Thirty to\n   forty lines. Thin sections stay thin.\n7. **Scrub and check.** Run the template's checklist. No email, account identifier or message\n   content survives into the report.\n8. **Publish** the report as a new revision of the product's pulse page, and read it back.\n9. **Surface.** Post the Headlines verbatim, the top Followup if it looks urgent, and where the report\n   lives. Never paste the whole report into chat.\n10. **Mention a reminder, lightly.** At setup, suggest once that the human keep a recurring reminder\n    to run `\u002Fak:product-pulse`; on the third or later ad hoc run, mention it once. A schedule may only\n    remind the human; it never starts a run, and this skill creates no reminder or schedule.\n\n## Hard gates\n\nGate: read-only everywhere. No write to the product, its database or any external system; a\nread-write database credential is refused at setup and never used if offered later.\n\nGate: no personal data in the report. No email, account identifier or message content, in any\nsection, the quality sample included.\n\nGate: no roadmap, strategy or tracker change. A request to act on the numbers is named as outside\nthis skill and left for a separate human-started run.\n\nGate: no invented numbers. A delta needs the prior equal window; without one the delta is omitted. A\nsource with no answer reads `no data`.\n\n| The thought | Why it is wrong | Do this instead |\n|---|---|---|\n| \"Give me write access and I can fix the counts at the source.\" | A pulse that mutates what it measures is not a read, and the credential outlives the run. | Refuse; offer a read-only connection or no database at all. |\n| \"Their emails would make the error list easier to debug.\" | The report is durable and shared; personal data in it is a leak. | Describe the error without identifiers; the native tool has the details. |\n| \"Signups dropped 30%, so I'll move onboarding to the top of the roadmap.\" | A report is not permission to change the roadmap. | Put it in Followups and stop. |\n| \"No prior window, but it's probably up about 10%.\" | An estimated delta reads as a measurement. | Omit the delta. |\n| \"This error count is high, I'll flag it red.\" | No thresholds were configured; a label is a judgment the reader did not ask for. | Present the number and let the reader judge. |\n\n## Outputs\n\n- A pulse report: a new revision of the product's pulse page, a `concept` page scoped to the product,\n  published through the knowledgebase adapter's `publishArtifact` under a `kb-document` placement.\n  Its revision history is the timeline of past pulses.\n- On setup, the pulse configuration, a `concept` page scoped to the product, published the same way.\n- The knowledgebase resolves both locations; this skill writes nothing into the working repository\n  (ruling `central-kb-owns-project-artifacts`).\n- In chat: the Headlines, the top Followup if urgent, and where the report lives.\n\n## Side effects\n\n`external-fetch`, `scratch-write`, `kb-draft`, `kb-publish`. No `workspace-write`, no\n`tracker-write`.\n\n`kb-publish` is a remote side effect. Its idempotency key derives from the run, the operation, the\npage's stable remote identity and the artifact's hash; the read-back is the record the publish\nreturns, read before the write and confirmed after it (`adapters\u002Frunner-contract\u002FCONTRACT.md`,\n\"Idempotency\"). A publish whose read-back cannot be performed is `failed`, never complete.\n\n## Stop conditions\n\n- `complete` — the report is published and read back, and its headlines are in chat.\n- `needs-input` — not started with `\u002Fak:product-pulse`, the knowledgebase cannot be read or written,\n  no signal source is configured or reachable, the window is unparseable, or a setup answer is\n  awaited.\n- `failed` — a publish's read-back cannot be performed.\n- `cancelled` — the human withdrew; nothing further is published.\n\n## Limits\n\n- Report length: thirty to forty lines (guidance).\n- Followups: three to five, each specific enough to act on (guidance).\n- Top errors: five, or the configured count (guidance).\n- Quality sample, when enabled: up to ten sessions (guidance).\n- Trailing buffer on the window: fifteen minutes (guidance).\n",{"name":5,"license":15,"metadata":24,"description":7},{"ak_catalog_id":5},"main","e0232821c4abbcd4338d6018c20fb70f1b74336b","skills\u002Fproduct-pulse",3,"https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit","https:\u002F\u002Fgithub.com\u002FPibomeister","www.eduardopicazoenriquez.com","2026-11-15 00:56:58",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add Pibomeister\u002Fagent-kit --skill product-pulse",{"tool":39,"command":40},"bunx","bunx skills add Pibomeister\u002Fagent-kit --skill product-pulse",{"tool":42,"command":43},"pnpm","pnpm dlx skills add Pibomeister\u002Fagent-kit --skill product-pulse","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Fskills\u002Fproduct-pulse\u002FSKILL.md",true,[47,53,59,63],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"assets\u002Freport-template.md","ASSET","资产文件",2294,"https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Fskills\u002Fproduct-pulse\u002Fassets\u002Freport-template.md",{"path":54,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":57,"sourceHtmlUrl":58},"references\u002Frun.md","REFERENCE","参考文档",3599,"https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Fskills\u002Fproduct-pulse\u002Freferences\u002Frun.md",{"path":60,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":61,"sourceHtmlUrl":62},"references\u002Fsetup.md",2744,"https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Fskills\u002Fproduct-pulse\u002Freferences\u002Fsetup.md",{"path":64,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":65,"sourceHtmlUrl":66},"skill.yaml",11401,"https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Fskills\u002Fproduct-pulse\u002Fskill.yaml",[68,75,80,85,90,95],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":19},"01M7D3JZP2026RGY6XX1MC3NQQ","alpha","Human-started command: it runs only when the human's message begins with `\u002Fak:alpha`. On any other request do not load or follow it; tell the human to type that command. Runs the alpha workflow when a human asks for it.","https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fvalid\u002Fminimal\u002Fskills\u002Falpha\u002FSKILL.md",[],[],{"skillNo":76,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M7D3JA593Q3RBYJXKVZ37Y2K","https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Finvalid\u002F25-article-tier-outside-vocabulary\u002Fskills\u002Falpha\u002FSKILL.md",[],[],{"skillNo":81,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7D3HQDDAZTAPGWCPGDPNVBQ","https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Finvalid\u002F24-article-without-triggers\u002Fskills\u002Falpha\u002FSKILL.md",[],[],{"skillNo":86,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M7D3HK074TK3ZWK1ZZNQVPB0","https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Finvalid\u002F23-enforced-by-without-predicate\u002Fskills\u002Falpha\u002FSKILL.md",[],[],{"skillNo":91,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7D3HGHQM5C5TBSXBCE2WFHW","https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Finvalid\u002F22-article-without-source-span\u002Fskills\u002Falpha\u002FSKILL.md",[],[],{"skillNo":96,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":99,"publishedAt":19},"01M7D3HDV6HMK2ZHRM0P368KNT","https:\u002F\u002Fgithub.com\u002FPibomeister\u002Fagent-kit\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Finvalid\u002F21-fragment-redeclares-id\u002Fskills\u002Falpha\u002FSKILL.md",[],[]]