[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6C1YPS6DHMHXM7J56M2BMFS":3,"skill-same:01M6C1YPS6DHMHXM7J56M2BMFS":58},{"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":32,"linkOnly":33,"markdownContent":34,"markdownBody":35,"frontMatter":36,"refName":37,"commitSha":38,"skillDirectory":39,"forksCount":40,"repositoryHtmlUrl":41,"repositoryHomepageUrl":6,"ownerHtmlUrl":42,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":43,"install":44,"resources":57},"01M6C1YPS6DHMHXM7J56M2BMFS","bypass-claims",null,"Bypass-claim evidence standards for any fast-path, short-circuit, LLM skip, template bypass, or cache hit claim in this repo.","jleechanorg\u002Fjleechan-skills","jleechanorg","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F221404966?v=4","01M6C1VJ99XG950X1DWPEBCBGX","01M6C1VJDK5YVDNT3362RT5WSB",6,"MIT","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fbypass-claims\u002FSKILL.md",[],[18,19,20,21,22,23,24,25,26,27,28,29,30,31],"bypass claims","claims","cache","fast","template","repo","short","evidence standards","evidence","standards","claim","circuit","path","for","2026-11-10 10:12:30",false,"---\nname: bypass-claims\ndescription: Bypass-claim evidence standards for any fast-path, short-circuit, LLM skip, template bypass, or cache hit claim in this repo.\n---\n\n# Bypass-claim evidence standards\n\n**Applies to**: any \"fast-path\", \"short-circuit\", \"LLM skip\", \"template bypass\", or \"cache hit\" claim in this repo.\n\n## Core rule\n\nA test that proves a code path **bypasses an expensive operation** is only valid if the input it feeds the system is **captured from the real client**, not reconstructed from server-side reference files the bypass itself consults.\n\nIf the test's input and the bypass's gate both derive from the same file, the test is tautological — it proves the helper works when handed its own reference input, not that production actually reaches the helper.\n\n## Required evidence for any bypass claim\n\n1. **Client-captured input** — the exact byte string the real client sends, captured via one of:\n   - Browser DevTools network tab export (HAR file)\n   - Playwright\u002FPuppeteer intercepted request body\n   - Firestore-persisted payload from a real production session\n   - JS constant extracted directly from the frontend source the browser serves (e.g. regex extract of `DEFAULT_X` from `campaign-wizard.js`)\n   - **NOT ALLOWED**: reading the same backend file the bypass's gate reads\n   - **NOT ALLOWED**: reconstructing the input from helper functions that share code with the gate\n2. **Latency delta proof** — measured wall-clock time of the request with the bypass vs. without. Bypass must be meaningfully faster (typically ≥5× for LLM skips).\n3. **Server log proof** — grep of the real server's stdout\u002Fstderr for the bypass's log line (e.g. `\"Using pre-generated Dragon Knight template opening\"`) AND absence of the expensive operation's log line (e.g. `\"generate_content\"`, `\"\u002Finteraction\u002Fstream\"` chunks).\n4. **Production trace match** — if the bypass claim concerns an existing production bug, fetch a real affected record from Firestore\u002Flogs and show its persisted input hash\u002Fbytes match what your test feeds.\n5. **Sync guard** — a CI-blocking unit test asserting the two (or more) sources of the bypass gate input stay byte-equal. Hash constants baked into `constants.py` MUST have a generator script AND a test that regenerates and compares.\n\n## Forbidden patterns\n\n- Loading the test description from `$PROJECT_ROOT\u002Fdata\u002F*_canonical_description.txt` and feeding it to an endpoint whose bypass hashes that same file → **circular provenance**\n- Calling the bypass helper directly in a unit test and claiming it proves \"the server bypasses the LLM\" → **unit test masquerading as pipeline test**\n- Measuring latency by calling the helper function in a loop → **does not cover routing, middleware, auth, prompt assembly, or any layer between ingress and the gate**\n- Any test whose PASS does not require the real HTTP handler to execute end-to-end → not a bypass claim test\n\n## Claim-class verdict rules\n\n| Evidence present | Verdict |\n|---|---|\n| Unit test of helper only | **Unit** — cannot claim bypass |\n| Helper + real HTTP call with server-derived input | **INSUFFICIENT** — circular provenance |\n| Helper + real HTTP call with client-captured input + latency delta + server log grep | **PASS** for bypass claim |\n| Above + sync guard test + production trace match | **PASS** for pipeline-bypass claim |\n\n## Writing the failing test first (RED)\n\nBefore fixing any bypass bug:\n\n1. Write the test that feeds **client-captured** input to the real local server (`.\u002Flocal.sh`).\n2. Assert the latency threshold AND the bypass log line AND the absence of the expensive-operation log.\n3. Run it — it MUST fail on the current broken code. If it passes on broken code, the test is not exercising the real path.\n4. Only then fix the root cause.\n5. Rerun — it MUST pass.\n6. Record both RED and GREEN evidence bundles with full absolute paths.\n\n## Sync guard pattern\n\nWhenever a bypass gate depends on two or more sources of truth being byte-identical (a JS constant and a backend file, a generated prompt and a baked-in SHA, etc.), the repo MUST include:\n\n- A regeneration script (e.g. `scripts\u002Fregenerate_dragon_knight_canonical.py`) that rebuilds the hash\u002Fconstant from the canonical source\n- A CI-enforced unit test that runs the regeneration in memory and asserts equality with the committed constant\n- A comment on the constant pointing to the regeneration script\n\nWithout the sync guard, any future edit to either source silently breaks the bypass — the same bug this skill exists to prevent.\n\n## Canonical frontend-source helper (Dragon Knight)\n\n`testing_utils\u002Fdragon_knight_frontend_source.py` exposes\n`get_frontend_default_dragon_knight_description()` — the **only** sanctioned\nway for any Dragon Knight bypass test to obtain the description field. It\nstatically parses `$PROJECT_ROOT\u002Ffrontend_v1\u002Fjs\u002Fcampaign-wizard.js` for the\n`DEFAULT_DRAGON_KNIGHT_DESCRIPTION` template literal and fails loudly if the\nshape changes. Any test that reads\n`$PROJECT_ROOT\u002Fdata\u002Fdragon_knight_canonical_description.txt` for its request\npayload is a circular-provenance violation and must be rewritten to use this\nhelper.\n\n## Checklist for any new fast-path \u002F bypass \u002F cache-hit claim\n\nBefore shipping an optimization that claims to skip an expensive operation on\na subset of inputs, the repo MUST grow the following five artifacts. The\nDragon Knight fast-path is the template; copy its structure.\n\n1. **Client-captured input helper** under `testing_utils\u002F` (mirror of\n   `testing_utils\u002Fdragon_knight_frontend_source.py`). The helper is the\n   single sanctioned way to obtain the \"real client bytes\" any bypass test\n   needs. It MUST NOT read the backend reference file the gate consults.\n2. **Sync-guard unit test** (mirror of\n   `$PROJECT_ROOT\u002Ftests\u002Ftest_dragon_knight_frontend_backend_sync.py`). Asserts\n   that running the gate's own prompt builder on the client-captured input\n   produces the baked-in SHA \u002F cache key. Wired into a job required by\n   branch protection so drift blocks merges.\n3. **Regeneration script** under `scripts\u002F` (mirror of\n   `scripts\u002Fregenerate_dragon_knight_canonical.py`). One command rebuilds\n   the constant (and any coupled backend data file) from the client helper.\n   Supports `--check` for a CI dry-run. Referenced by a comment directly\n   above the constant in code.\n4. **Pipeline integration test with fatal LLM spy** (mirror of\n   `$PROJECT_ROOT\u002Ftests\u002Ftest_dragon_knight_pipeline_evidence.py`). Drives the\n   real server function (not the helper) with client-captured input,\n   patches every LLM entrypoint with an `AssertionError`-raising spy, and\n   asserts zero spy hits on the canonical path and ≥1 on a negative\n   control. Evidence bundle allocated via `setUpModule()` with atomic\n   `os.replace()` for the `latest` symlink so parallel runs don't race.\n5. **Real-server UI test** (mirror of\n   `testing_ui\u002Ftest_dragon_knight_fastpath_real_frontend.py`). Boots\n   gunicorn, POSTs client-captured bytes through the real frontend helper,\n   asserts the fast-path marker appears in server stdout and no LLM API\n   markers do, AND records wall-clock latency delta vs. a non-canonical\n   control. CI path filter MUST include every file the gate reads (JS\n   constant, backend canonical file, prompt builder, template module).\n\nIf any of these five artifacts is missing, the bypass claim is INSUFFICIENT\nregardless of how many helper-level tests pass.\n\n## Prior incident (PR #6137)\n\nDragon Knight fast-path claimed to skip `get_initial_story()` LLM call for default campaigns. Three separate \"evidence\" tests (commits `351d3c24d`, `12960e46b`, `285d89699`) all PASSED. Reality: preview server took 32s on real dragon-knight creation. Root cause: tests loaded `description` from `$PROJECT_ROOT\u002Fdata\u002Fdragon_knight_canonical_description.txt` (18k chars) — the same file the bypass's SHA gate hashes — while the actual frontend `campaign-wizard.js` ships a 41k-char version. The hash could never match in production; the test could never detect it.\n\nFix requires: (a) either sync frontend byte-equal to backend canonical OR drop the SHA gate in favor of `campaign_type + custom_options` matching, AND (b) a real-server UI test that posts the frontend default string AND (c) a sync guard CI test.\n","# Bypass-claim evidence standards\n\n**Applies to**: any \"fast-path\", \"short-circuit\", \"LLM skip\", \"template bypass\", or \"cache hit\" claim in this repo.\n\n## Core rule\n\nA test that proves a code path **bypasses an expensive operation** is only valid if the input it feeds the system is **captured from the real client**, not reconstructed from server-side reference files the bypass itself consults.\n\nIf the test's input and the bypass's gate both derive from the same file, the test is tautological — it proves the helper works when handed its own reference input, not that production actually reaches the helper.\n\n## Required evidence for any bypass claim\n\n1. **Client-captured input** — the exact byte string the real client sends, captured via one of:\n   - Browser DevTools network tab export (HAR file)\n   - Playwright\u002FPuppeteer intercepted request body\n   - Firestore-persisted payload from a real production session\n   - JS constant extracted directly from the frontend source the browser serves (e.g. regex extract of `DEFAULT_X` from `campaign-wizard.js`)\n   - **NOT ALLOWED**: reading the same backend file the bypass's gate reads\n   - **NOT ALLOWED**: reconstructing the input from helper functions that share code with the gate\n2. **Latency delta proof** — measured wall-clock time of the request with the bypass vs. without. Bypass must be meaningfully faster (typically ≥5× for LLM skips).\n3. **Server log proof** — grep of the real server's stdout\u002Fstderr for the bypass's log line (e.g. `\"Using pre-generated Dragon Knight template opening\"`) AND absence of the expensive operation's log line (e.g. `\"generate_content\"`, `\"\u002Finteraction\u002Fstream\"` chunks).\n4. **Production trace match** — if the bypass claim concerns an existing production bug, fetch a real affected record from Firestore\u002Flogs and show its persisted input hash\u002Fbytes match what your test feeds.\n5. **Sync guard** — a CI-blocking unit test asserting the two (or more) sources of the bypass gate input stay byte-equal. Hash constants baked into `constants.py` MUST have a generator script AND a test that regenerates and compares.\n\n## Forbidden patterns\n\n- Loading the test description from `$PROJECT_ROOT\u002Fdata\u002F*_canonical_description.txt` and feeding it to an endpoint whose bypass hashes that same file → **circular provenance**\n- Calling the bypass helper directly in a unit test and claiming it proves \"the server bypasses the LLM\" → **unit test masquerading as pipeline test**\n- Measuring latency by calling the helper function in a loop → **does not cover routing, middleware, auth, prompt assembly, or any layer between ingress and the gate**\n- Any test whose PASS does not require the real HTTP handler to execute end-to-end → not a bypass claim test\n\n## Claim-class verdict rules\n\n| Evidence present | Verdict |\n|---|---|\n| Unit test of helper only | **Unit** — cannot claim bypass |\n| Helper + real HTTP call with server-derived input | **INSUFFICIENT** — circular provenance |\n| Helper + real HTTP call with client-captured input + latency delta + server log grep | **PASS** for bypass claim |\n| Above + sync guard test + production trace match | **PASS** for pipeline-bypass claim |\n\n## Writing the failing test first (RED)\n\nBefore fixing any bypass bug:\n\n1. Write the test that feeds **client-captured** input to the real local server (`.\u002Flocal.sh`).\n2. Assert the latency threshold AND the bypass log line AND the absence of the expensive-operation log.\n3. Run it — it MUST fail on the current broken code. If it passes on broken code, the test is not exercising the real path.\n4. Only then fix the root cause.\n5. Rerun — it MUST pass.\n6. Record both RED and GREEN evidence bundles with full absolute paths.\n\n## Sync guard pattern\n\nWhenever a bypass gate depends on two or more sources of truth being byte-identical (a JS constant and a backend file, a generated prompt and a baked-in SHA, etc.), the repo MUST include:\n\n- A regeneration script (e.g. `scripts\u002Fregenerate_dragon_knight_canonical.py`) that rebuilds the hash\u002Fconstant from the canonical source\n- A CI-enforced unit test that runs the regeneration in memory and asserts equality with the committed constant\n- A comment on the constant pointing to the regeneration script\n\nWithout the sync guard, any future edit to either source silently breaks the bypass — the same bug this skill exists to prevent.\n\n## Canonical frontend-source helper (Dragon Knight)\n\n`testing_utils\u002Fdragon_knight_frontend_source.py` exposes\n`get_frontend_default_dragon_knight_description()` — the **only** sanctioned\nway for any Dragon Knight bypass test to obtain the description field. It\nstatically parses `$PROJECT_ROOT\u002Ffrontend_v1\u002Fjs\u002Fcampaign-wizard.js` for the\n`DEFAULT_DRAGON_KNIGHT_DESCRIPTION` template literal and fails loudly if the\nshape changes. Any test that reads\n`$PROJECT_ROOT\u002Fdata\u002Fdragon_knight_canonical_description.txt` for its request\npayload is a circular-provenance violation and must be rewritten to use this\nhelper.\n\n## Checklist for any new fast-path \u002F bypass \u002F cache-hit claim\n\nBefore shipping an optimization that claims to skip an expensive operation on\na subset of inputs, the repo MUST grow the following five artifacts. The\nDragon Knight fast-path is the template; copy its structure.\n\n1. **Client-captured input helper** under `testing_utils\u002F` (mirror of\n   `testing_utils\u002Fdragon_knight_frontend_source.py`). The helper is the\n   single sanctioned way to obtain the \"real client bytes\" any bypass test\n   needs. It MUST NOT read the backend reference file the gate consults.\n2. **Sync-guard unit test** (mirror of\n   `$PROJECT_ROOT\u002Ftests\u002Ftest_dragon_knight_frontend_backend_sync.py`). Asserts\n   that running the gate's own prompt builder on the client-captured input\n   produces the baked-in SHA \u002F cache key. Wired into a job required by\n   branch protection so drift blocks merges.\n3. **Regeneration script** under `scripts\u002F` (mirror of\n   `scripts\u002Fregenerate_dragon_knight_canonical.py`). One command rebuilds\n   the constant (and any coupled backend data file) from the client helper.\n   Supports `--check` for a CI dry-run. Referenced by a comment directly\n   above the constant in code.\n4. **Pipeline integration test with fatal LLM spy** (mirror of\n   `$PROJECT_ROOT\u002Ftests\u002Ftest_dragon_knight_pipeline_evidence.py`). Drives the\n   real server function (not the helper) with client-captured input,\n   patches every LLM entrypoint with an `AssertionError`-raising spy, and\n   asserts zero spy hits on the canonical path and ≥1 on a negative\n   control. Evidence bundle allocated via `setUpModule()` with atomic\n   `os.replace()` for the `latest` symlink so parallel runs don't race.\n5. **Real-server UI test** (mirror of\n   `testing_ui\u002Ftest_dragon_knight_fastpath_real_frontend.py`). Boots\n   gunicorn, POSTs client-captured bytes through the real frontend helper,\n   asserts the fast-path marker appears in server stdout and no LLM API\n   markers do, AND records wall-clock latency delta vs. a non-canonical\n   control. CI path filter MUST include every file the gate reads (JS\n   constant, backend canonical file, prompt builder, template module).\n\nIf any of these five artifacts is missing, the bypass claim is INSUFFICIENT\nregardless of how many helper-level tests pass.\n\n## Prior incident (PR #6137)\n\nDragon Knight fast-path claimed to skip `get_initial_story()` LLM call for default campaigns. Three separate \"evidence\" tests (commits `351d3c24d`, `12960e46b`, `285d89699`) all PASSED. Reality: preview server took 32s on real dragon-knight creation. Root cause: tests loaded `description` from `$PROJECT_ROOT\u002Fdata\u002Fdragon_knight_canonical_description.txt` (18k chars) — the same file the bypass's SHA gate hashes — while the actual frontend `campaign-wizard.js` ships a 41k-char version. The hash could never match in production; the test could never detect it.\n\nFix requires: (a) either sync frontend byte-equal to backend canonical OR drop the SHA gate in favor of `campaign_type + custom_options` matching, AND (b) a real-server UI test that posts the frontend default string AND (c) a sync guard CI test.\n",{"name":5,"description":7},"main","42ccd02cbf5e8e47839dc3eab6fc754dc6b339cf",".claude\u002Fskills\u002Fbypass-claims",0,"https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills","https:\u002F\u002Fgithub.com\u002Fjleechanorg","2026-11-11 22:19:36",{"source":8,"skillName":5,"commands":45,"prompt":55,"zipAvailable":56},[46,49,52],{"tool":47,"command":48},"npx","npx skills add jleechanorg\u002Fjleechan-skills --skill bypass-claims",{"tool":50,"command":51},"bunx","bunx skills add jleechanorg\u002Fjleechan-skills --skill bypass-claims",{"tool":53,"command":54},"pnpm","pnpm dlx skills add jleechanorg\u002Fjleechan-skills --skill bypass-claims","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fbypass-claims\u002FSKILL.md",true,[],[59,77,98,112,126,142],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":32},"01M716Y5MTJ6GYJZ2ATXCJK8G4","instinct","Use when the user invokes \u002Finstinct or \u002Fin, or asks to \"message Instinct\", \"ask Instinct\", \"tell Instinct\", \"check Instinct\", or read replies from the configured peer assistant. iMessage first, WhatsApp Web as fallback.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Finstinct\u002FSKILL.md",[],[61,66,67,68,69,70,71,72,73,74,75,76],"user","when","replies","check","use","identity","whatsapp","gmail","read","message","ask",{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":88,"publishedAt":32},"01M716Y4HF7FVVMJSA2CCVN5VF","hr","Compatibility alias for history-resume; produce the same fresh-state, provenance-aware resume briefing.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fhr\u002FSKILL.md",[83],{"categoryNo":84,"categoryName":85,"taxonomyType":86,"taxonomyTypeDesc":87},"01JBSEEDCAT000000000000038","人力资源","PROFESSION","适用职业",[79,89,90,91,31,92,93,94,95,96,97],"resume","compatibility","provenance","history","produce","alias","briefing","state","fresh",{"skillNo":99,"name":100,"nameZh":6,"description":101,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":102,"categories":103,"keywords":105,"publishedAt":32},"01M716Y1FFNFKFJYYKF8HQA6T8","history-resume","Resume briefing for the current repository branch or worktree using fresh state and admitted history sources.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fhistory-resume\u002FSKILL.md",[104],{"categoryNo":84,"categoryName":85,"taxonomyType":86,"taxonomyTypeDesc":87},[92,89,106,107,66,108,95,109,31,110,111,96,97],"history resume","branch","analysis","repository","worktree","sources",{"skillNo":113,"name":114,"nameZh":6,"description":115,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":116,"categories":117,"keywords":118,"publishedAt":32},"01M716XWXJ8EPHK4RH9T14TR85","grok","Alias for the canonical grokbot browser conversation workflow; \u002Fgrok talks to the Grok bot the same way \u002Fdot talks to Dot.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fgrok\u002FSKILL.md",[],[114,119,120,121,31,122,94,123,124,125],"dot","bot","workflow","canonical","conversation","browser","grokbot",{"skillNo":127,"name":128,"nameZh":6,"description":129,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":130,"categories":131,"keywords":132,"publishedAt":32},"01M70N9Q07DTE0EW5KFXMMMSJ0","codexa-bench","Compare `codexa` and normal `codex` on a fair split of comparable work by default; use `--strict` for duplicate runs of the exact same task.","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fcodexa-bench\u002FSKILL.md",[],[133,134,135,70,136,31,137,138,139,140,141],"codexa bench","work","split","default","duplicate","task","compare","normal","codex",{"skillNo":143,"name":144,"nameZh":6,"description":145,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":146,"licenseFileUrl":147,"sourceHtmlUrl":148,"categories":149,"keywords":161,"publishedAt":32},"01M6C24S2X60SP8BBFGAEKPEGW","frontend-design","Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, artifacts, posters, or applications (examples include websites, landing pages, dashboards, React components, HTML\u002FCSS layouts, or when styling\u002Fbeautifying any web UI). Generates creative, polished code and UI design that avoids generic AI aesthetics.","Apache-2.0","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ffrontend-design\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fjleechanorg\u002Fjleechan-skills\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ffrontend-design\u002FSKILL.md",[150,155,158],{"categoryNo":151,"categoryName":152,"taxonomyType":153,"taxonomyTypeDesc":154},"01JBSEEDCAT000000000000009","前端开发","SCENARIO","使用场景",{"categoryNo":156,"categoryName":157,"taxonomyType":153,"taxonomyTypeDesc":154},"01JBSEEDCAT000000000000015","视觉设计",{"categoryNo":159,"categoryName":160,"taxonomyType":86,"taxonomyTypeDesc":87},"01JBSEEDCAT000000000000032","设计师",[162,163,164,70,165,166,167,66,168,169,170,171,172,173,174,175,176,177,178,179,180,181,182,183,184,185,67,186],"frontend","design","frontend design","generic","ui design","react","html","components","code","artifacts","ui","aesthetics","Production","web components","landing","applications","quality","web","css","create","examples","with","react components","web ui","pages"]