[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M72B2FAASCK528Q16S8RSF5W":3,"skill-same:01M72B2FAASCK528Q16S8RSF5W":220},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M72B2FAASCK528Q16S8RSF5W","aurora-design-coherence",null,"audit, redesign, benchmark, and govern auroramind B2B application interfaces for ChatGPT or Codex. Use for new apps or existing Nexus, NKS, B DATA Studio, Nexus Pocket, Silio, AIS\u002FCIS, MIA, Compta-IA, SocialGenie, or other Auroramind products when the task involves UI\u002FUX architecture, screen hierarchy, information overload, developer-facing UI leakage, Playwright-based visual inspection, Refero\u002FMobbin design research, premium B2B patterns, target mockups, shadcn\u002FAurora component mapping, Codex implementation specs, or before\u002Fafter visual QA without changing business logic.","syl2042\u002FAurora_SR_method_codex_pack","syl2042","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F78928611?v=4","01M72B1YGG8QB8JYDPY2XJBS1H","01M72B1YH6N787N2PA3A1HBFKW",1,"MIT","https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002FSKILL.md",[],[],"2026-11-10 18:39:20",false,"---\nname: aurora-design-coherence\ndescription: audit, redesign, benchmark, and govern auroramind B2B application interfaces for ChatGPT or Codex. Use for new apps or existing Nexus, NKS, B DATA Studio, Nexus Pocket, Silio, AIS\u002FCIS, MIA, Compta-IA, SocialGenie, or other Auroramind products when the task involves UI\u002FUX architecture, screen hierarchy, information overload, developer-facing UI leakage, Playwright-based visual inspection, Refero\u002FMobbin design research, premium B2B patterns, target mockups, shadcn\u002FAurora component mapping, Codex implementation specs, or before\u002Fafter visual QA without changing business logic.\n---\n\n# Aurora Design Coherence V2.1 — Design Director\n\nAct as the **AI Design Director** for Auroramind products. Own the product-interface analysis, UX\u002FUI decisions, information hierarchy, benchmark research, target design, and implementation specification. The human product owner is the **final validator**. Codex is the **implementation agent** and must not invent the redesign after validation.\n\nThe objective is not to make screens prettier. Convert technically capable applications into premium B2B products that expose the right information, at the right time, to the right persona.\n\n## Non-negotiable role split\n\n- **Design Director**: understand the product, inspect evidence, decide the target UX\u002FUI, generate the target visual, and write deterministic implementation specs.\n- **Human validator**: approve, reject, or request changes to the target. Do not require the human to use Figma, CSS, shadcn, or any design tool.\n- **Codex implementer**: implement the approved target and spec. Do not reinterpret the information architecture or visual hierarchy unless the approved spec is technically impossible.\n- **No implementation before visual approval** in redesign mode unless the user explicitly waives this gate.\n\n## Operating modes\n\n1. **existing-app redesign** — preferred for current Auroramind apps. Inspect code + live UI, research current patterns, redesign screen-by-screen, render target, obtain human approval, then hand off to Codex.\n2. **new-app design** — define product jobs, navigation, screen inventory, component system, target screens, then prepare Codex tasks before implementation.\n3. **audit only** — inspect and report without redesigning or modifying code.\n4. **design review** — critique screenshots, prototypes, generated UI, or a proposed target.\n5. **migration planning** — turn an approved redesign into phased implementation tasks.\n6. **visual QA** — compare approved target with implemented UI and issue pass\u002Fminor-fix\u002Ffail findings.\n7. **kit governance** — extract validated recurring patterns into Aurora shared components.\n\nUse the smallest mode that satisfies the request. For a request such as “rework this app\u002Fpage\u002FUI”, default to **existing-app redesign**.\n\n# Existing-app redesign workflow\n\nFollow these gates in order. Do not collapse them into a one-shot rewrite.\n\n## Gate 0 — Evidence and scope\n\n1. Identify repository\u002Fref and target environment.\n2. Inspect the codebase sufficiently to understand the feature, routes, business objects, existing components, themes, and permissions.\n3. Inspect the **rendered application** with the browser when available. Prefer Playwright\u002Fbrowser evidence over code-only inference.\n4. Capture representative current-state screenshots before making design decisions.\n5. Record evidence confidence precisely: `rendered`, `measured`, `code`, `design-reference`, or `inferred`.\n\nLoad `references\u002Fbrowser-evidence.md` for authenticated browser and screenshot rules.\n\nFor the user's normal Auroramind setup, prefer the already available **Codex App → Remote SSH → OVH repository** path plus the configured **Playwright MCP\u002Fbrowser**. If the deployed\u002Fdev application is reachable over HTTPS, do not introduce an SSH tunnel merely to inspect it.\n\n## Gate 1 — Product intent before components\n\nFor every target screen, establish:\n\n- persona;\n- job to be done;\n- primary question or decision;\n- primary action;\n- secondary actions;\n- frequency of use;\n- business criticality;\n- information required to complete the job;\n- information that is useful only for experts\u002Fadmins\u002Fdevelopers.\n\nLoad `references\u002Fproduct-intent.md`.\n\nDo not start with “which shadcn component should replace this card?”. First decide whether the information or interaction should exist at this level at all.\n\n## Gate 2 — Information triage\n\nInventory all meaningful information and actions visible on the screen. Classify each item:\n\n- `PROMOTE` — should gain visual priority;\n- `KEEP` — correct level and location;\n- `GROUP` — retain but consolidate with related information;\n- `MOVE` — retain but move to another area\u002Fscreen\u002Fdrawer;\n- `HIDE` — keep available through progressive disclosure or diagnostics;\n- `REMOVE` — no user value in this context.\n\nLoad `references\u002Finformation-triage.md`.\n\n**Availability in the backend is never sufficient justification for visibility in the UI.**\n\n## Gate 3 — Current design research and decision lock\n\nBefore a material redesign, research current real-product patterns when a design research source is available. External references inform the decision; they do not own it.\n\nPreferred sources:\n\n1. **Refero MCP** — primary source for the initial Auroramind workflow.\n2. **Mobbin MCP** — complementary breadth when available.\n3. Existing approved Aurora patterns.\n\nFor **in-app B2B product UI**, research in this order:\n\n1. screens for concrete hierarchy, content, states, and interaction;\n2. similar screens when one candidate strongly matches;\n3. flows when the task spans multiple steps or state transitions;\n4. styles only after product architecture is understood, when visual craft needs external direction.\n\nFor marketing\u002Fart-direction work, styles may come first.\n\nLoad `references\u002Fdesign-research.md` and `references\u002Fresearch-contract.md`.\n\nFor each material target screen:\n\n- state a precise research question;\n- search at least 5 relevant candidates when the corpus supports it;\n- inspect and shortlist the strongest candidates;\n- assign each retained reference a bounded role;\n- create a **REFERENCE LOCK** before target generation;\n- create a **DECISION LEDGER** tracing every major decision to product evidence, Aurora rules, or external research;\n- apply the anti-averaging rule: do not blend strong references into generic SaaS styling;\n- never copy another product's branding or visual identity.\n\nIf external research is unavailable, explicitly mark `current benchmark unavailable`, use approved Aurora patterns + product evidence, and still produce a Decision Ledger.\n\n## Gate 4 — Product\u002FUX audit, visual craft audit, and redesign specification\n\nProduce a `SCREEN_REDESIGN_SPEC.md` using `templates\u002FSCREEN_REDESIGN_SPEC.md`.\n\nThe spec must contain:\n\n- current evidence;\n- separate Product\u002FUX Audit and Visual Craft Audit;\n- persona\u002Fjob\u002Fdecision\u002Faction;\n- identified UX\u002Fproduct problems;\n- information triage table;\n- selected reference patterns;\n- research question and query trail;\n- Reference Lock;\n- Decision Ledger;\n- target information architecture;\n- exact navigation behavior;\n- layout and density rules;\n- current component → target component mapping;\n- progressive disclosure levels;\n- technical\u002Fadmin information placement;\n- responsive behavior;\n- empty\u002Floading\u002Ferror\u002Fpermission\u002Fprocessing states;\n- target files likely affected;\n- non-goals and protected business logic;\n- acceptance criteria.\n\nA redesign spec is incomplete if it only says “simplify”, “modernize”, “make premium”, or names components without deciding information hierarchy.\n\nLoad `references\u002Fscreen-redesign.md`.\n\n## Gate 5 — Generate the target visual\n\nA target visual is **required before implementation** for a material redesign.\n\nGenerate it from:\n\n- current screenshot(s);\n- approved information triage;\n- target information architecture;\n- selected design-reference patterns;\n- Aurora\u002Fshadcn component constraints;\n- existing theme\u002Fbrand behavior.\n\nUse the best available visual renderer. Prefer direct image generation for the first concept-validation loop. Magic Patterns, Stitch, Figma Make, or another design renderer may be used only as an AI-operated renderer; do not require the human validator to design manually.\n\nThe target visual must represent the written spec. Do not create an unrelated “pretty SaaS” mockup.\n\nLoad `references\u002Ftarget-generation.md`.\n\n## Gate 6 — Human validation\n\nPresent the human with at minimum:\n\n- `CURRENT` screenshot;\n- `TARGET` visual;\n- concise explanation of the main structural changes.\n\nThe human can approve or request changes. Iterate the **target**, not production code, until approved.\n\nUse the exact state label `TARGET APPROVED` only after explicit human validation.\n\nLoad `references\u002Fhuman-validation.md`.\n\n## Gate 7 — Codex handoff\n\nAfter `TARGET APPROVED`, produce `CODEX_UI_TASK.md` from `templates\u002FCODEX_UI_TASK_V2.md`.\n\nCodex receives:\n\n- approved target visual;\n- `SCREEN_REDESIGN_SPEC.md`;\n- target files;\n- component mapping;\n- allowed changes;\n- forbidden changes;\n- tests;\n- visual acceptance criteria.\n\nCodex is implementing an approved design contract, not being asked to “make it nicer”.\n\nLoad `references\u002Fcodex-handoff-v2.md`.\n\n## Gate 8 — Before\u002Ftarget\u002Fafter QA\n\nAfter implementation, capture the same screen using comparable:\n\n- route;\n- persona\u002Fpermissions;\n- viewport;\n- representative data\u002Fstate.\n\nCompare:\n\n`CURRENT → TARGET → IMPLEMENTED`\n\nReturn one verdict:\n\n- `PASS` — materially matches target and product intent;\n- `MINOR FIX` — target preserved, localized issues remain;\n- `FAIL` — important hierarchy, disclosure, navigation, or target fidelity is wrong.\n\nGive file\u002Fcomponent-specific corrections. Do not accept a technically working screen that violates the approved product design.\n\nLoad `references\u002Fvisual-qa-v2.md`.\n\n## Gate 9 — Pattern capture\n\nWhen the human approves a reusable solution, determine whether it should become an `AURORA APPROVED PATTERN` or shared component. Prefer reuse over redesigning the same problem independently in another Auroramind app.\n\nLoad `references\u002Fapproved-patterns.md`.\n\n# New-app workflow\n\nFor a new app, perform the same reasoning without a current screenshot:\n\n1. define personas, jobs, business objects, risk and maturity;\n2. define navigation and screen inventory;\n3. research current patterns for the key screen families;\n4. create screen blueprints and information hierarchy;\n5. generate target visuals for critical screens;\n6. obtain human approval;\n7. produce incremental Codex implementation tasks;\n8. run visual QA against approved targets.\n\nReuse the existing templates `APP_DESIGN_BRIEF.md`, `APP_DESIGN_SYSTEM.md`, `APP_SCREEN_BLUEPRINT.md`, and `CODEX_NEW_APP_UI_TASKS.md` when relevant.\n\n# Premium B2B rules\n\nPremium B2B means operational clarity, not decorative luxury.\n\n- Prefer clear hierarchy, calm density, predictable actions, readable data, restrained motion, and explicit state.\n- Optimize for the user's decision and task, not for displaying every technical capability.\n- Avoid card grids as the default composition. Cards group coherent decisions; they are not generic containers for every piece of data.\n- Expose technical details through progressive disclosure unless the target persona genuinely needs them.\n- Separate normal work, contextual detail, advanced configuration, and diagnostics.\n- Preserve user-selectable themes. Do not impose a new global color identity unless explicitly requested.\n- Prefer existing Aurora wrappers for identity-defining patterns and shadcn\u002FRadix as primitives.\n- Use Tremor\u002FKibo only through Aurora wrappers when wrappers exist.\n- Decorative gradients, glow, glassmorphism, fake KPI dashboards, random icons, marketing heroes in operational screens, and motion without workflow value are AI-slop signals.\n\nLoad `references\u002Fpremium-b2b-selection.md`, `references\u002Fcomponent-tiers.md`, and `references\u002Fforbidden-patterns.md` when making component or visual decisions.\n\n# AI product rules\n\nFor RAG\u002Fagent\u002Fdata products, never equate technical explainability with permanent visual exposure. Evidence, sources, traces, provider\u002Fmodel metadata, costs, IDs, and logs have different user value by persona.\n\nLoad `references\u002Fdesign-skills\u002Fai-product-patterns.md` and apply `references\u002Finformation-triage.md` before deciding what is visible by default.\n\n# Existing repo checks\n\nInspect when relevant:\n\n- frontend framework\u002Frouting;\n- `shared\u002Fui`, manifests, `components\u002Fui`;\n- page\u002Flayout routes;\n- `package.json`, Tailwind\u002Fglobal CSS\u002Ftheme providers;\n- shell\u002Fsidebar\u002Ftopbar;\n- headers, tables, cards, badges, filters, drawers;\n- AI\u002Fsource\u002Ftrace\u002Fevidence UI;\n- direct Tremor\u002FKibo imports;\n- hardcoded semantic colors;\n- developer\u002Fadmin\u002Fdebug information exposed to normal users.\n\nIf local repository files are available, optionally use the existing read-only inventory scripts:\n\n```bash\npython scripts\u002Fscan_ui_inventory.py \u003Crepo-root> --output aurora-ui-inventory.json --markdown aurora-ui-inventory.md\npython scripts\u002Fscore_design_coherence.py aurora-ui-inventory.json --output AURORA_DESIGN_AUDIT_DRAFT.md\n```\n\nTreat script scores as codebase evidence only; they cannot judge rendered hierarchy or product intent.\n\n# Protected boundaries\n\nUnless explicitly requested, do not change:\n\n- business logic;\n- API contracts;\n- database models;\n- authentication\u002Fauthorization;\n- production routing;\n- data semantics;\n- user-selectable theme behavior.\n\nA UI redesign may reorganize presentation and interaction around existing capabilities, but it must not silently invent or remove business behavior.\n\n# Core references\n\nLoad only what the current task needs:\n\n- `references\u002Fproduct-intent.md` — persona\u002Fjob\u002Fdecision\u002Faction model.\n- `references\u002Finformation-triage.md` — visibility and progressive-disclosure decisions.\n- `references\u002Fbrowser-evidence.md` — Playwright\u002Fbrowser evidence and authentication handling.\n- `references\u002Fdesign-research.md` — live Refero\u002FMobbin research protocol.\n- `references\u002Fresearch-contract.md` — Research Question, Reference Lock, Decision Ledger, anti-averaging, and dual audit rules.\n- `references\u002Fscreen-redesign.md` — screen redesign decision framework.\n- `references\u002Ftarget-generation.md` — visual target generation and fidelity rules.\n- `references\u002Fhuman-validation.md` — final human validation gate.\n- `references\u002Fcodex-handoff-v2.md` — deterministic implementation handoff.\n- `references\u002Fvisual-qa-v2.md` — current\u002Ftarget\u002Fimplemented comparison.\n- `references\u002Fapproved-patterns.md` — reusable Aurora pattern capture.\n- `references\u002Fcomponent-tiers.md` — shared Aurora components.\n- `references\u002Fpremium-b2b-selection.md` — component and density guidance.\n- `references\u002Fforbidden-patterns.md` — anti-patterns.\n- `references\u002Fdesign-skills\u002Fai-product-patterns.md` — AI\u002FRAG\u002Fagent-specific UI.\n- `references\u002Fdesign-skills\u002Fmobile-b2b.md` — mobile constraints.\n- `references\u002Fmigration-rules.md` — safe migration rules.\n\n# Templates\n\nUse as needed:\n\n- `templates\u002FSCREEN_REDESIGN_SPEC.md`\n- `templates\u002FCODEX_UI_TASK_V2.md`\n- `templates\u002FVISUAL_QA_V2.md`\n- `templates\u002FAPP_DESIGN_BRIEF.md`\n- `templates\u002FAPP_DESIGN_SYSTEM.md`\n- `templates\u002FAPP_SCREEN_BLUEPRINT.md`\n- `templates\u002FCODEX_NEW_APP_UI_TASKS.md`\n","# Aurora Design Coherence V2.1 — Design Director\n\nAct as the **AI Design Director** for Auroramind products. Own the product-interface analysis, UX\u002FUI decisions, information hierarchy, benchmark research, target design, and implementation specification. The human product owner is the **final validator**. Codex is the **implementation agent** and must not invent the redesign after validation.\n\nThe objective is not to make screens prettier. Convert technically capable applications into premium B2B products that expose the right information, at the right time, to the right persona.\n\n## Non-negotiable role split\n\n- **Design Director**: understand the product, inspect evidence, decide the target UX\u002FUI, generate the target visual, and write deterministic implementation specs.\n- **Human validator**: approve, reject, or request changes to the target. Do not require the human to use Figma, CSS, shadcn, or any design tool.\n- **Codex implementer**: implement the approved target and spec. Do not reinterpret the information architecture or visual hierarchy unless the approved spec is technically impossible.\n- **No implementation before visual approval** in redesign mode unless the user explicitly waives this gate.\n\n## Operating modes\n\n1. **existing-app redesign** — preferred for current Auroramind apps. Inspect code + live UI, research current patterns, redesign screen-by-screen, render target, obtain human approval, then hand off to Codex.\n2. **new-app design** — define product jobs, navigation, screen inventory, component system, target screens, then prepare Codex tasks before implementation.\n3. **audit only** — inspect and report without redesigning or modifying code.\n4. **design review** — critique screenshots, prototypes, generated UI, or a proposed target.\n5. **migration planning** — turn an approved redesign into phased implementation tasks.\n6. **visual QA** — compare approved target with implemented UI and issue pass\u002Fminor-fix\u002Ffail findings.\n7. **kit governance** — extract validated recurring patterns into Aurora shared components.\n\nUse the smallest mode that satisfies the request. For a request such as “rework this app\u002Fpage\u002FUI”, default to **existing-app redesign**.\n\n# Existing-app redesign workflow\n\nFollow these gates in order. Do not collapse them into a one-shot rewrite.\n\n## Gate 0 — Evidence and scope\n\n1. Identify repository\u002Fref and target environment.\n2. Inspect the codebase sufficiently to understand the feature, routes, business objects, existing components, themes, and permissions.\n3. Inspect the **rendered application** with the browser when available. Prefer Playwright\u002Fbrowser evidence over code-only inference.\n4. Capture representative current-state screenshots before making design decisions.\n5. Record evidence confidence precisely: `rendered`, `measured`, `code`, `design-reference`, or `inferred`.\n\nLoad `references\u002Fbrowser-evidence.md` for authenticated browser and screenshot rules.\n\nFor the user's normal Auroramind setup, prefer the already available **Codex App → Remote SSH → OVH repository** path plus the configured **Playwright MCP\u002Fbrowser**. If the deployed\u002Fdev application is reachable over HTTPS, do not introduce an SSH tunnel merely to inspect it.\n\n## Gate 1 — Product intent before components\n\nFor every target screen, establish:\n\n- persona;\n- job to be done;\n- primary question or decision;\n- primary action;\n- secondary actions;\n- frequency of use;\n- business criticality;\n- information required to complete the job;\n- information that is useful only for experts\u002Fadmins\u002Fdevelopers.\n\nLoad `references\u002Fproduct-intent.md`.\n\nDo not start with “which shadcn component should replace this card?”. First decide whether the information or interaction should exist at this level at all.\n\n## Gate 2 — Information triage\n\nInventory all meaningful information and actions visible on the screen. Classify each item:\n\n- `PROMOTE` — should gain visual priority;\n- `KEEP` — correct level and location;\n- `GROUP` — retain but consolidate with related information;\n- `MOVE` — retain but move to another area\u002Fscreen\u002Fdrawer;\n- `HIDE` — keep available through progressive disclosure or diagnostics;\n- `REMOVE` — no user value in this context.\n\nLoad `references\u002Finformation-triage.md`.\n\n**Availability in the backend is never sufficient justification for visibility in the UI.**\n\n## Gate 3 — Current design research and decision lock\n\nBefore a material redesign, research current real-product patterns when a design research source is available. External references inform the decision; they do not own it.\n\nPreferred sources:\n\n1. **Refero MCP** — primary source for the initial Auroramind workflow.\n2. **Mobbin MCP** — complementary breadth when available.\n3. Existing approved Aurora patterns.\n\nFor **in-app B2B product UI**, research in this order:\n\n1. screens for concrete hierarchy, content, states, and interaction;\n2. similar screens when one candidate strongly matches;\n3. flows when the task spans multiple steps or state transitions;\n4. styles only after product architecture is understood, when visual craft needs external direction.\n\nFor marketing\u002Fart-direction work, styles may come first.\n\nLoad `references\u002Fdesign-research.md` and `references\u002Fresearch-contract.md`.\n\nFor each material target screen:\n\n- state a precise research question;\n- search at least 5 relevant candidates when the corpus supports it;\n- inspect and shortlist the strongest candidates;\n- assign each retained reference a bounded role;\n- create a **REFERENCE LOCK** before target generation;\n- create a **DECISION LEDGER** tracing every major decision to product evidence, Aurora rules, or external research;\n- apply the anti-averaging rule: do not blend strong references into generic SaaS styling;\n- never copy another product's branding or visual identity.\n\nIf external research is unavailable, explicitly mark `current benchmark unavailable`, use approved Aurora patterns + product evidence, and still produce a Decision Ledger.\n\n## Gate 4 — Product\u002FUX audit, visual craft audit, and redesign specification\n\nProduce a `SCREEN_REDESIGN_SPEC.md` using `templates\u002FSCREEN_REDESIGN_SPEC.md`.\n\nThe spec must contain:\n\n- current evidence;\n- separate Product\u002FUX Audit and Visual Craft Audit;\n- persona\u002Fjob\u002Fdecision\u002Faction;\n- identified UX\u002Fproduct problems;\n- information triage table;\n- selected reference patterns;\n- research question and query trail;\n- Reference Lock;\n- Decision Ledger;\n- target information architecture;\n- exact navigation behavior;\n- layout and density rules;\n- current component → target component mapping;\n- progressive disclosure levels;\n- technical\u002Fadmin information placement;\n- responsive behavior;\n- empty\u002Floading\u002Ferror\u002Fpermission\u002Fprocessing states;\n- target files likely affected;\n- non-goals and protected business logic;\n- acceptance criteria.\n\nA redesign spec is incomplete if it only says “simplify”, “modernize”, “make premium”, or names components without deciding information hierarchy.\n\nLoad `references\u002Fscreen-redesign.md`.\n\n## Gate 5 — Generate the target visual\n\nA target visual is **required before implementation** for a material redesign.\n\nGenerate it from:\n\n- current screenshot(s);\n- approved information triage;\n- target information architecture;\n- selected design-reference patterns;\n- Aurora\u002Fshadcn component constraints;\n- existing theme\u002Fbrand behavior.\n\nUse the best available visual renderer. Prefer direct image generation for the first concept-validation loop. Magic Patterns, Stitch, Figma Make, or another design renderer may be used only as an AI-operated renderer; do not require the human validator to design manually.\n\nThe target visual must represent the written spec. Do not create an unrelated “pretty SaaS” mockup.\n\nLoad `references\u002Ftarget-generation.md`.\n\n## Gate 6 — Human validation\n\nPresent the human with at minimum:\n\n- `CURRENT` screenshot;\n- `TARGET` visual;\n- concise explanation of the main structural changes.\n\nThe human can approve or request changes. Iterate the **target**, not production code, until approved.\n\nUse the exact state label `TARGET APPROVED` only after explicit human validation.\n\nLoad `references\u002Fhuman-validation.md`.\n\n## Gate 7 — Codex handoff\n\nAfter `TARGET APPROVED`, produce `CODEX_UI_TASK.md` from `templates\u002FCODEX_UI_TASK_V2.md`.\n\nCodex receives:\n\n- approved target visual;\n- `SCREEN_REDESIGN_SPEC.md`;\n- target files;\n- component mapping;\n- allowed changes;\n- forbidden changes;\n- tests;\n- visual acceptance criteria.\n\nCodex is implementing an approved design contract, not being asked to “make it nicer”.\n\nLoad `references\u002Fcodex-handoff-v2.md`.\n\n## Gate 8 — Before\u002Ftarget\u002Fafter QA\n\nAfter implementation, capture the same screen using comparable:\n\n- route;\n- persona\u002Fpermissions;\n- viewport;\n- representative data\u002Fstate.\n\nCompare:\n\n`CURRENT → TARGET → IMPLEMENTED`\n\nReturn one verdict:\n\n- `PASS` — materially matches target and product intent;\n- `MINOR FIX` — target preserved, localized issues remain;\n- `FAIL` — important hierarchy, disclosure, navigation, or target fidelity is wrong.\n\nGive file\u002Fcomponent-specific corrections. Do not accept a technically working screen that violates the approved product design.\n\nLoad `references\u002Fvisual-qa-v2.md`.\n\n## Gate 9 — Pattern capture\n\nWhen the human approves a reusable solution, determine whether it should become an `AURORA APPROVED PATTERN` or shared component. Prefer reuse over redesigning the same problem independently in another Auroramind app.\n\nLoad `references\u002Fapproved-patterns.md`.\n\n# New-app workflow\n\nFor a new app, perform the same reasoning without a current screenshot:\n\n1. define personas, jobs, business objects, risk and maturity;\n2. define navigation and screen inventory;\n3. research current patterns for the key screen families;\n4. create screen blueprints and information hierarchy;\n5. generate target visuals for critical screens;\n6. obtain human approval;\n7. produce incremental Codex implementation tasks;\n8. run visual QA against approved targets.\n\nReuse the existing templates `APP_DESIGN_BRIEF.md`, `APP_DESIGN_SYSTEM.md`, `APP_SCREEN_BLUEPRINT.md`, and `CODEX_NEW_APP_UI_TASKS.md` when relevant.\n\n# Premium B2B rules\n\nPremium B2B means operational clarity, not decorative luxury.\n\n- Prefer clear hierarchy, calm density, predictable actions, readable data, restrained motion, and explicit state.\n- Optimize for the user's decision and task, not for displaying every technical capability.\n- Avoid card grids as the default composition. Cards group coherent decisions; they are not generic containers for every piece of data.\n- Expose technical details through progressive disclosure unless the target persona genuinely needs them.\n- Separate normal work, contextual detail, advanced configuration, and diagnostics.\n- Preserve user-selectable themes. Do not impose a new global color identity unless explicitly requested.\n- Prefer existing Aurora wrappers for identity-defining patterns and shadcn\u002FRadix as primitives.\n- Use Tremor\u002FKibo only through Aurora wrappers when wrappers exist.\n- Decorative gradients, glow, glassmorphism, fake KPI dashboards, random icons, marketing heroes in operational screens, and motion without workflow value are AI-slop signals.\n\nLoad `references\u002Fpremium-b2b-selection.md`, `references\u002Fcomponent-tiers.md`, and `references\u002Fforbidden-patterns.md` when making component or visual decisions.\n\n# AI product rules\n\nFor RAG\u002Fagent\u002Fdata products, never equate technical explainability with permanent visual exposure. Evidence, sources, traces, provider\u002Fmodel metadata, costs, IDs, and logs have different user value by persona.\n\nLoad `references\u002Fdesign-skills\u002Fai-product-patterns.md` and apply `references\u002Finformation-triage.md` before deciding what is visible by default.\n\n# Existing repo checks\n\nInspect when relevant:\n\n- frontend framework\u002Frouting;\n- `shared\u002Fui`, manifests, `components\u002Fui`;\n- page\u002Flayout routes;\n- `package.json`, Tailwind\u002Fglobal CSS\u002Ftheme providers;\n- shell\u002Fsidebar\u002Ftopbar;\n- headers, tables, cards, badges, filters, drawers;\n- AI\u002Fsource\u002Ftrace\u002Fevidence UI;\n- direct Tremor\u002FKibo imports;\n- hardcoded semantic colors;\n- developer\u002Fadmin\u002Fdebug information exposed to normal users.\n\nIf local repository files are available, optionally use the existing read-only inventory scripts:\n\n```bash\npython scripts\u002Fscan_ui_inventory.py \u003Crepo-root> --output aurora-ui-inventory.json --markdown aurora-ui-inventory.md\npython scripts\u002Fscore_design_coherence.py aurora-ui-inventory.json --output AURORA_DESIGN_AUDIT_DRAFT.md\n```\n\nTreat script scores as codebase evidence only; they cannot judge rendered hierarchy or product intent.\n\n# Protected boundaries\n\nUnless explicitly requested, do not change:\n\n- business logic;\n- API contracts;\n- database models;\n- authentication\u002Fauthorization;\n- production routing;\n- data semantics;\n- user-selectable theme behavior.\n\nA UI redesign may reorganize presentation and interaction around existing capabilities, but it must not silently invent or remove business behavior.\n\n# Core references\n\nLoad only what the current task needs:\n\n- `references\u002Fproduct-intent.md` — persona\u002Fjob\u002Fdecision\u002Faction model.\n- `references\u002Finformation-triage.md` — visibility and progressive-disclosure decisions.\n- `references\u002Fbrowser-evidence.md` — Playwright\u002Fbrowser evidence and authentication handling.\n- `references\u002Fdesign-research.md` — live Refero\u002FMobbin research protocol.\n- `references\u002Fresearch-contract.md` — Research Question, Reference Lock, Decision Ledger, anti-averaging, and dual audit rules.\n- `references\u002Fscreen-redesign.md` — screen redesign decision framework.\n- `references\u002Ftarget-generation.md` — visual target generation and fidelity rules.\n- `references\u002Fhuman-validation.md` — final human validation gate.\n- `references\u002Fcodex-handoff-v2.md` — deterministic implementation handoff.\n- `references\u002Fvisual-qa-v2.md` — current\u002Ftarget\u002Fimplemented comparison.\n- `references\u002Fapproved-patterns.md` — reusable Aurora pattern capture.\n- `references\u002Fcomponent-tiers.md` — shared Aurora components.\n- `references\u002Fpremium-b2b-selection.md` — component and density guidance.\n- `references\u002Fforbidden-patterns.md` — anti-patterns.\n- `references\u002Fdesign-skills\u002Fai-product-patterns.md` — AI\u002FRAG\u002Fagent-specific UI.\n- `references\u002Fdesign-skills\u002Fmobile-b2b.md` — mobile constraints.\n- `references\u002Fmigration-rules.md` — safe migration rules.\n\n# Templates\n\nUse as needed:\n\n- `templates\u002FSCREEN_REDESIGN_SPEC.md`\n- `templates\u002FCODEX_UI_TASK_V2.md`\n- `templates\u002FVISUAL_QA_V2.md`\n- `templates\u002FAPP_DESIGN_BRIEF.md`\n- `templates\u002FAPP_DESIGN_SYSTEM.md`\n- `templates\u002FAPP_SCREEN_BLUEPRINT.md`\n- `templates\u002FCODEX_NEW_APP_UI_TASKS.md`\n",{"name":5,"description":7},"main","596a681742e664dc9fdbbdc26783d62bba3f77f0","skills\u002Faurora-design-coherence",0,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack","https:\u002F\u002Fgithub.com\u002Fsyl2042","2026-09-25 01:24:24",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add syl2042\u002FAurora_SR_method_codex_pack --skill aurora-design-coherence",{"tool":36,"command":37},"bunx","bunx skills add syl2042\u002FAurora_SR_method_codex_pack --skill aurora-design-coherence",{"tool":39,"command":40},"pnpm","pnpm dlx skills add syl2042\u002FAurora_SR_method_codex_pack --skill aurora-design-coherence","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002FSKILL.md",true,[44,50,56,60,64,68,72,76,80,84,88,92,96,100,104,108,112,116,120,124,128,132,136,140,144,148,152,156,160,166,170,174,178,182,186,190,194,198,202,206,210,214],{"path":45,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":48,"sourceHtmlUrl":49},"agents\u002Fopenai.yaml","REFERENCE","参考文档",387,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Fagents\u002Fopenai.yaml",{"path":51,"resourceType":52,"resourceTypeDesc":53,"sizeBytes":54,"sourceHtmlUrl":55},"assets\u002Ficon.svg","ASSET","资产文件",1210,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Fassets\u002Ficon.svg",{"path":57,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":58,"sourceHtmlUrl":59},"references\u002Fapproved-patterns.md",1028,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fapproved-patterns.md",{"path":61,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":62,"sourceHtmlUrl":63},"references\u002Faudit-method.md",2269,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Faudit-method.md",{"path":65,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":66,"sourceHtmlUrl":67},"references\u002Fbrowser-evidence.md",1898,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fbrowser-evidence.md",{"path":69,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":70,"sourceHtmlUrl":71},"references\u002Fcodex-handoff-v2.md",1193,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fcodex-handoff-v2.md",{"path":73,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":74,"sourceHtmlUrl":75},"references\u002Fcomponent-tiers.md",1843,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fcomponent-tiers.md",{"path":77,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":78,"sourceHtmlUrl":79},"references\u002Fdesign-research.md",2922,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fdesign-research.md",{"path":81,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":82,"sourceHtmlUrl":83},"references\u002Fdesign-skills\u002Fai-product-patterns.md",2301,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fdesign-skills\u002Fai-product-patterns.md",{"path":85,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":86,"sourceHtmlUrl":87},"references\u002Fdesign-skills\u002Fdesign-director.md",2001,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fdesign-skills\u002Fdesign-director.md",{"path":89,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":90,"sourceHtmlUrl":91},"references\u002Fdesign-skills\u002Fdesign-system-authoring.md",1418,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fdesign-skills\u002Fdesign-system-authoring.md",{"path":93,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":94,"sourceHtmlUrl":95},"references\u002Fdesign-skills\u002Fmigration-codex.md",1444,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fdesign-skills\u002Fmigration-codex.md",{"path":97,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":98,"sourceHtmlUrl":99},"references\u002Fdesign-skills\u002Fmobile-b2b.md",1095,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fdesign-skills\u002Fmobile-b2b.md",{"path":101,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":102,"sourceHtmlUrl":103},"references\u002Fdesign-skills\u002Fproduct-cockpit.md",1701,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fdesign-skills\u002Fproduct-cockpit.md",{"path":105,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":106,"sourceHtmlUrl":107},"references\u002Fdesign-skills\u002Fvisual-qa.md",1300,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fdesign-skills\u002Fvisual-qa.md",{"path":109,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":110,"sourceHtmlUrl":111},"references\u002Fforbidden-patterns.md",1757,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fforbidden-patterns.md",{"path":113,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":114,"sourceHtmlUrl":115},"references\u002Fgeneric-b2b-pack-roadmap.md",1808,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fgeneric-b2b-pack-roadmap.md",{"path":117,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":118,"sourceHtmlUrl":119},"references\u002Fhuman-validation.md",988,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fhuman-validation.md",{"path":121,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":122,"sourceHtmlUrl":123},"references\u002Finformation-triage.md",2010,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Finformation-triage.md",{"path":125,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":126,"sourceHtmlUrl":127},"references\u002Fmigration-rules.md",1917,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fmigration-rules.md",{"path":129,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":130,"sourceHtmlUrl":131},"references\u002Fpremium-b2b-selection.md",1878,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fpremium-b2b-selection.md",{"path":133,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":134,"sourceHtmlUrl":135},"references\u002Fproduct-intent.md",1385,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fproduct-intent.md",{"path":137,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":138,"sourceHtmlUrl":139},"references\u002Freport-templates.md",2642,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Freport-templates.md",{"path":141,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":142,"sourceHtmlUrl":143},"references\u002Fresearch-contract.md",5109,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fresearch-contract.md",{"path":145,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":146,"sourceHtmlUrl":147},"references\u002Fscreen-redesign.md",2058,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fscreen-redesign.md",{"path":149,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":150,"sourceHtmlUrl":151},"references\u002Ftarget-generation.md",2661,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Ftarget-generation.md",{"path":153,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":154,"sourceHtmlUrl":155},"references\u002Ftremor-kibo-policy.md",1575,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Ftremor-kibo-policy.md",{"path":157,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":158,"sourceHtmlUrl":159},"references\u002Fvisual-qa-v2.md",1356,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Freferences\u002Fvisual-qa-v2.md",{"path":161,"resourceType":162,"resourceTypeDesc":163,"sizeBytes":164,"sourceHtmlUrl":165},"scripts\u002Fscan_ui_inventory.py","SCRIPT","脚本",8667,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Fscripts\u002Fscan_ui_inventory.py",{"path":167,"resourceType":162,"resourceTypeDesc":163,"sizeBytes":168,"sourceHtmlUrl":169},"scripts\u002Fscore_design_coherence.py",9522,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Fscripts\u002Fscore_design_coherence.py",{"path":171,"resourceType":162,"resourceTypeDesc":163,"sizeBytes":172,"sourceHtmlUrl":173},"scripts\u002Fvalidate_redesign_spec.py",1892,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Fscripts\u002Fvalidate_redesign_spec.py",{"path":175,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":176,"sourceHtmlUrl":177},"templates\u002FAPP_DESIGN_BRIEF.md",1509,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FAPP_DESIGN_BRIEF.md",{"path":179,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":180,"sourceHtmlUrl":181},"templates\u002FAPP_DESIGN_SYSTEM.md",2606,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FAPP_DESIGN_SYSTEM.md",{"path":183,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":184,"sourceHtmlUrl":185},"templates\u002FAPP_SCREEN_BLUEPRINT.md",1638,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FAPP_SCREEN_BLUEPRINT.md",{"path":187,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":188,"sourceHtmlUrl":189},"templates\u002FCODEX_NEW_APP_UI_TASKS.md",2458,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FCODEX_NEW_APP_UI_TASKS.md",{"path":191,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":192,"sourceHtmlUrl":193},"templates\u002FCODEX_UI_TASK_V2.md",1576,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FCODEX_UI_TASK_V2.md",{"path":195,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":196,"sourceHtmlUrl":197},"templates\u002FCODEX_UI_TASK.md",1366,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FCODEX_UI_TASK.md",{"path":199,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":200,"sourceHtmlUrl":201},"templates\u002FDESIGN_REVIEW.md",890,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FDESIGN_REVIEW.md",{"path":203,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":204,"sourceHtmlUrl":205},"templates\u002FSCREEN_REDESIGN_SPEC.md",2701,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FSCREEN_REDESIGN_SPEC.md",{"path":207,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":208,"sourceHtmlUrl":209},"templates\u002FVISUAL_QA_CHECKLIST.md",1471,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FVISUAL_QA_CHECKLIST.md",{"path":211,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":212,"sourceHtmlUrl":213},"templates\u002FVISUAL_QA_V2.md",679,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002Ftemplates\u002FVISUAL_QA_V2.md",{"path":215,"resourceType":216,"resourceTypeDesc":217,"sizeBytes":218,"sourceHtmlUrl":219},"VERSION","OTHER","其他",6,"https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills\u002Faurora-design-coherence\u002FVERSION",[221,228,235,242,249,256],{"skillNo":222,"name":223,"nameZh":6,"description":224,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":225,"categories":226,"keywords":227,"publishedAt":18},"01M72B2D9GHRGV7YNMQAFC76D9","aurora-ui-visual-qa","Validate a significant UI change visually or through an existing browser path.","https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills-method\u002Faurora-ui-visual-qa\u002FSKILL.md",[],[],{"skillNo":229,"name":230,"nameZh":6,"description":231,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":232,"categories":233,"keywords":234,"publishedAt":18},"01M72B2BEDVGP2NJWGWWAEHHJB","aurora-to-prd","Frame explicit product discovery when a request is not implementable as stated.","https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills-method\u002Faurora-to-prd\u002FSKILL.md",[],[],{"skillNo":236,"name":237,"nameZh":6,"description":238,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":239,"categories":240,"keywords":241,"publishedAt":18},"01M72B29DY27SZEYAR40343MYE","aurora-lot-runner","Execute a validated multi-lot pass, a long resume, or a bounded autonomous phase.","https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills-method\u002Faurora-lot-runner\u002FSKILL.md",[],[],{"skillNo":243,"name":244,"nameZh":6,"description":245,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":246,"categories":247,"keywords":248,"publishedAt":18},"01M72B282BQXNWTNMRXKBJZSBA","aurora-domain-skill-factory","Create local domain skills during project bootstrap or when an explicitly needed domain skill is absent.","https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills-method\u002Faurora-domain-skill-factory\u002FSKILL.md",[],[],{"skillNo":250,"name":251,"nameZh":6,"description":252,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":253,"categories":254,"keywords":255,"publishedAt":18},"01M72B25XCGK3CY4W2RAMC5S7Q","aurora-diagnose","Explain or correct an observed bug, runtime error, failed test, or integration problem.","https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills-method\u002Faurora-diagnose\u002FSKILL.md",[],[],{"skillNo":257,"name":258,"nameZh":6,"description":259,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":260,"categories":261,"keywords":262,"publishedAt":18},"01M72B23RFC3FNHKDYW07RP80Z","aurora-architecture-check","Review a durable structural decision: schema, migration, dependency, external integration, or orchestration boundary.","https:\u002F\u002Fgithub.com\u002Fsyl2042\u002FAurora_SR_method_codex_pack\u002Fblob\u002Fmain\u002Fskills-method\u002Faurora-architecture-check\u002FSKILL.md",[],[]]