[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7B1JMEMJWMF1KMN9R4NHF3E":3,"skill-same:01M7B1JMEMJWMF1KMN9R4NHF3E":49},{"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":27,"commitSha":28,"skillDirectory":29,"forksCount":14,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":32,"ownerDescription":33,"updatedAt":34,"install":35,"resources":48},"01M7B1JMEMJWMF1KMN9R4NHF3E","red-pen",null,"Turn a raw design red-pen — a pile of screenshots plus dictated critique of the generated scenarios — into deduped, correctly-routed, harvested changes to the AXP design harness. Use when the user dumps a critique of the scenario builds and wants it captured and actioned without losing anything, over-harvesting, or trusting the wrong source. The harvest half of the design-quality loop; pairs with product-loop (the generate half).","torkelo\u002Fscenes2","torkelo","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F10999?v=4","01M7B1G6RRGVK2TZGTJ2813VJ4","Torkel Ödegaard","01M7B1G6SAYVTDC3NWMBPX16TB",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Ftorkelo\u002Fscenes2\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fred-pen\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: red-pen\ndescription: Turn a raw design red-pen — a pile of screenshots plus dictated critique of the generated scenarios — into deduped, correctly-routed, harvested changes to the AXP design harness. Use when the user dumps a critique of the scenario builds and wants it captured and actioned without losing anything, over-harvesting, or trusting the wrong source. The harvest half of the design-quality loop; pairs with product-loop (the generate half).\nmetadata:\n  generated: hatch@v1.2.0\n  source: .hatch\u002F_skills\u002Fred-pen\u002FSKILL.md\n---\n\n# red-pen\n\nThe intake for a human critique of the generated scenarios, and the counterpart to `\u002Fproduct-loop`:\nproduct-loop generates and judges; red-pen turns the critique into routed changes to the harness in\n`product\u002F`. Read `product\u002FREADME.md` for the homes and invariants first. Run this pipeline in\norder; do not jump to editing.\n\n## 1. Parse the dump faithfully\n\nThe input is voice dictation (filler, run-ons, transcription errors — \"insect\" → \"inset\",\nprofanity as emphasis) plus screenshots that reference specific `scenario\u002FvN`. Parse the _intended\nmeaning_, map every screenshot to its scenario and version, and when a point is genuinely ambiguous\n**ask rather than guess**.\n\n## 2. Verify — trust neither the dictation nor the builders\n\nGround each claim against the actual render (screenshot the story on `:6006` — never spawn a server,\nsee `no-spawning-storybook` — or read the attached crop). Both the critique and the builders' own\n`caughtFixed` reports have been wrong. Mark what you could not verify (behind a tab, scroll\nbehaviour) rather than asserting it.\n\n## 3. Dedupe to recurring patterns — the highest-value step\n\nCollapse the notes into a small set of patterns grouped by _underlying cause_, not by scenario. The\nrecurrence is the insight (inconsistency is the meta-defect): \"eyes zigzag \u002F nothing lines up \u002F on\ntop of itself\" is one row-alignment finding, not fifteen.\n\n## 4. For each pattern, ask the higher-level question\n\nDo not just log the point-fix. For each cluster, name the _foundational_ change that would stop the\nwhole class (\"what do we author so a hard black border is never reached for\") and only then the\nspecific fix beneath it. Generalisation first.\n\n## 5. Route each finding to exactly one home\n\nPer `product\u002FREADME.md`: a composition rule → the matching `docs\u002Fdesign\u002F*` reference; a\ncomponent-agnostic fundamental → `docs\u002Fdesign\u002Ffoundations.md`; a non-checkable judgment → `docs\u002Fdesign\u002Ftaste.md`; a shipped\nvalue bug → `component-specs\u002FCOMPONENT-FIXES.md`; a cross-cutting state\u002Fbehaviour →\n`component-specs\u002FCONTRACTS.md`; a component to build → a spec + `BUILD-QUEUE.md`. Never restate a\ncaptured value (one home). Separate genuinely-new findings from confirmations of existing rules\n(confirmations add priority evidence, not new entries).\n\n## 6. Flag tensions — never silently overwrite a rule\n\nWhen a finding conflicts with an existing rule, or is a judgment call, surface it as a decision for\nthe user rather than editing over it. Do not over-harvest: a scenario-specific one-off that is\nalready an instance of a general rule gets no entry of its own.\n\n## 7. Output, then execute\n\nFirst a review artifact — (a) recurring patterns, (b) new rules by home, (c) confirmations,\n(d) tensions needing a decision, (e) where you'd push back. After the user's steer (or straight away\nfor the unambiguous ones), apply the edits, Prettier-clean them, and report **where each finding\nlanded**. Finally, note which new rules the next `product-loop` round should be able to test —\nclosing the loop back to generation.\n\n## Guardrails\n\nDedupe before routing · one home per value · verify, don't assume · generalise over point-fix ·\nflag, don't overwrite · preserve the user's voice and intent · don't over-harvest.\n","# red-pen\n\nThe intake for a human critique of the generated scenarios, and the counterpart to `\u002Fproduct-loop`:\nproduct-loop generates and judges; red-pen turns the critique into routed changes to the harness in\n`product\u002F`. Read `product\u002FREADME.md` for the homes and invariants first. Run this pipeline in\norder; do not jump to editing.\n\n## 1. Parse the dump faithfully\n\nThe input is voice dictation (filler, run-ons, transcription errors — \"insect\" → \"inset\",\nprofanity as emphasis) plus screenshots that reference specific `scenario\u002FvN`. Parse the _intended\nmeaning_, map every screenshot to its scenario and version, and when a point is genuinely ambiguous\n**ask rather than guess**.\n\n## 2. Verify — trust neither the dictation nor the builders\n\nGround each claim against the actual render (screenshot the story on `:6006` — never spawn a server,\nsee `no-spawning-storybook` — or read the attached crop). Both the critique and the builders' own\n`caughtFixed` reports have been wrong. Mark what you could not verify (behind a tab, scroll\nbehaviour) rather than asserting it.\n\n## 3. Dedupe to recurring patterns — the highest-value step\n\nCollapse the notes into a small set of patterns grouped by _underlying cause_, not by scenario. The\nrecurrence is the insight (inconsistency is the meta-defect): \"eyes zigzag \u002F nothing lines up \u002F on\ntop of itself\" is one row-alignment finding, not fifteen.\n\n## 4. For each pattern, ask the higher-level question\n\nDo not just log the point-fix. For each cluster, name the _foundational_ change that would stop the\nwhole class (\"what do we author so a hard black border is never reached for\") and only then the\nspecific fix beneath it. Generalisation first.\n\n## 5. Route each finding to exactly one home\n\nPer `product\u002FREADME.md`: a composition rule → the matching `docs\u002Fdesign\u002F*` reference; a\ncomponent-agnostic fundamental → `docs\u002Fdesign\u002Ffoundations.md`; a non-checkable judgment → `docs\u002Fdesign\u002Ftaste.md`; a shipped\nvalue bug → `component-specs\u002FCOMPONENT-FIXES.md`; a cross-cutting state\u002Fbehaviour →\n`component-specs\u002FCONTRACTS.md`; a component to build → a spec + `BUILD-QUEUE.md`. Never restate a\ncaptured value (one home). Separate genuinely-new findings from confirmations of existing rules\n(confirmations add priority evidence, not new entries).\n\n## 6. Flag tensions — never silently overwrite a rule\n\nWhen a finding conflicts with an existing rule, or is a judgment call, surface it as a decision for\nthe user rather than editing over it. Do not over-harvest: a scenario-specific one-off that is\nalready an instance of a general rule gets no entry of its own.\n\n## 7. Output, then execute\n\nFirst a review artifact — (a) recurring patterns, (b) new rules by home, (c) confirmations,\n(d) tensions needing a decision, (e) where you'd push back. After the user's steer (or straight away\nfor the unambiguous ones), apply the edits, Prettier-clean them, and report **where each finding\nlanded**. Finally, note which new rules the next `product-loop` round should be able to test —\nclosing the loop back to generation.\n\n## Guardrails\n\nDedupe before routing · one home per value · verify, don't assume · generalise over point-fix ·\nflag, don't overwrite · preserve the user's voice and intent · don't over-harvest.\n",{"name":5,"metadata":24,"description":7},{"source":25,"generated":26},".hatch\u002F_skills\u002Fred-pen\u002FSKILL.md","hatch@v1.2.0","main","9b85ba995139627c15a6ef1cec1daec64659cf75",".agents\u002Fskills\u002Fred-pen","https:\u002F\u002Fgithub.com\u002Ftorkelo\u002Fscenes2","https:\u002F\u002Fgithub.com\u002Ftorkelo","https:\u002F\u002Fgrafana.com","Grafana Labs Co-Founder, Grafana Creator & product manager & lead architect.  ","2026-11-14 19:30:54",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":47},[37,40,43],{"tool":38,"command":39},"npx","npx skills add torkelo\u002Fscenes2 --skill red-pen",{"tool":41,"command":42},"bunx","bunx skills add torkelo\u002Fscenes2 --skill red-pen",{"tool":44,"command":45},"pnpm","pnpm dlx skills add torkelo\u002Fscenes2 --skill red-pen","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Ftorkelo\u002Fscenes2\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fred-pen\u002FSKILL.md",true,[],[50,57,64,69,74,81],{"skillNo":51,"name":52,"nameZh":6,"description":53,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":19},"01M7B1QT196EFM12F7737ZFBJ3","validate-plugin","validate a Grafana plugin using the official Grafana plugin validator","https:\u002F\u002Fgithub.com\u002Ftorkelo\u002Fscenes2\u002Fblob\u002Fmain\u002Fapps\u002Fdemo-app\u002F.codex\u002Fskills\u002Fvalidate-plugin\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M7B1QQGHG8JJ9MK4BY2V8879","build-plugin","build a Grafana plugin using the standard build process (frontend and backend if applicable)","https:\u002F\u002Fgithub.com\u002Ftorkelo\u002Fscenes2\u002Fblob\u002Fmain\u002Fapps\u002Fdemo-app\u002F.codex\u002Fskills\u002Fbuild-plugin\u002FSKILL.md",[],[],{"skillNo":65,"name":52,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M7B1QPP9RN7E1KNMRK9D8V6M","https:\u002F\u002Fgithub.com\u002Ftorkelo\u002Fscenes2\u002Fblob\u002Fmain\u002Fapps\u002Fdemo-app\u002F.claude\u002Fskills\u002Fvalidate-plugin\u002FSKILL.md",[],[],{"skillNo":70,"name":59,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M7B1QN3V9SQ7H6WXNZDMQ57Q","https:\u002F\u002Fgithub.com\u002Ftorkelo\u002Fscenes2\u002Fblob\u002Fmain\u002Fapps\u002Fdemo-app\u002F.claude\u002Fskills\u002Fbuild-plugin\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M7B1QJEW06858M7GC4EM46B5","stop-slop","Remove AI writing patterns from prose. Use when drafting, editing, or reviewing text to eliminate predictable AI tells.","https:\u002F\u002Fgithub.com\u002Ftorkelo\u002Fscenes2\u002Fblob\u002Fmain\u002F.hatch\u002F_skills\u002Fstop-slop\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7B1QGEPPZQQNRJRXTQVZAKC","snag-fixer","Lightweight agent for quick fixes, tweaks, and small improvements. Optimized for single-task focused work without bloating context. Used by snag workflow for sequential fix processing.","https:\u002F\u002Fgithub.com\u002Ftorkelo\u002Fscenes2\u002Fblob\u002Fmain\u002F.hatch\u002F_skills\u002Fsnag-fixer\u002FSKILL.md",[],[]]