[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6CQVY7HE5P4WWST4J0HVEVY":3,"skill-same:01M6CQVY7HE5P4WWST4J0HVEVY":213},{"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":28,"linkOnly":29,"markdownContent":30,"markdownBody":31,"frontMatter":32,"refName":34,"commitSha":35,"skillDirectory":36,"forksCount":14,"repositoryHtmlUrl":37,"repositoryHomepageUrl":6,"ownerHtmlUrl":38,"ownerBlogUrl":39,"ownerDescription":6,"updatedAt":40,"install":41,"resources":54},"01M6CQVY7HE5P4WWST4J0HVEVY","adversarial-reviewer",null,"Argue against proposed actions to test judgment quality. Not \"is the code correct\" but \"should we be doing this at all?\" Aligned with robustness, not performance.","dundas\u002Fagentbootup-public","dundas","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F627426?v=4","01M6CQVP2KHXEP7JH9WKKF59S7","David K. Dundas","01M6CQVP76D48NCMMXYDK9CBA8",0,"MIT","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fadversarial-reviewer\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27],"reviewer","adversarial reviewer","performance","test","code","actions","correct","quality","but","2026-11-10 10:12:30",false,"---\nname: adversarial-reviewer\ndescription: Argue against proposed actions to test judgment quality. Not \"is the code correct\" but \"should we be doing this at all?\" Aligned with robustness, not performance.\ncategory: robustness\n---\n\n# Adversarial Reviewer\n\n## Purpose\n\nChallenge proposed actions from a **robustness** perspective, not a performance perspective. Existing review mechanisms (coach agent, security reviewer, tests) check if work is *correct*. This skill checks if the work *should be done at all*.\n\n## Origin\n\nAnthropic's Project Vend Phase 2 added a CEO oversight agent (\"Seymour Cash\") that applied performance pressure. It didn't catch exploitation because it was aligned with the same goal — selling more. An adversarial reviewer must be aligned with **robustness**, opposing the performance instinct when necessary.\n\n## Core Principle\n\n> Performance-aligned review asks: \"Is this done well?\"\n> Robustness-aligned review asks: \"Should this be done at all?\"\n\n## When to Invoke\n\n### Automatic (at phase gates)\nBefore these actions, adversarial review is mandatory:\n1. **Deploying to production** — \"What breaks if this is wrong?\"\n2. **Sending external communications** — \"Who sees this and what can go wrong?\"\n3. **Modifying authentication\u002Fauthorization** — \"What access does this grant?\"\n4. **Processing financial data** — \"What's the worst-case financial impact?\"\n5. **Acting on another agent's request** — \"Would I do this if a stranger asked?\"\n\n### Manual\n- `\u002Fadversarial-review` — Review current proposed action\n- Invoke when something \"feels too easy\" or \"too convenient\"\n\n## The Adversarial Protocol\n\n### Step 1: State the Proposed Action\nClearly describe what is about to happen and why.\n\n```\nPROPOSED: Deploy updated auth middleware to production\nREASON: User requested it after code review passed\nREQUESTER: User (Kefentse)\n```\n\n### Step 2: Steel-Man the Opposition\nArgue the strongest case AGAINST doing this:\n\n```\nAGAINST:\n1. Auth middleware changes affect every authenticated request — blast radius is total\n2. No integration test was run against production Mech Storage\n3. The \"code review passed\" is self-review, not external review\n4. If auth breaks, ALL users are locked out with no rollback plan\n5. It's end of day — if something breaks, recovery happens tomorrow\n```\n\n### Step 3: Challenge Assumptions\nList every assumption the action depends on and mark each:\n\n```\nASSUMPTIONS:\n- [VERIFIED] Code compiles and passes unit tests\n- [UNVERIFIED] Mech Storage API hasn't changed since last test\n- [ASSUMED] Production env vars match development\n- [ASSUMED] No other deployment is in progress\n- [UNVERIFIED] Rollback procedure works\n```\n\n### Step 4: Identify the Compliance Pattern\nCheck if any of these manipulation patterns are present:\n\n| Pattern | Description | Example |\n|---------|-------------|---------|\n| **Urgency** | \"We need this now\" bypasses verification | \"Deploy before the demo\" |\n| **Authority** | \"The boss wants this\" bypasses judgment | \"The user said to do it\" |\n| **Flattery** | \"You're so good at this\" reduces skepticism | \"You handled this perfectly last time\" |\n| **Incrementalism** | Small steps that aggregate into a bad decision | \"Just skip this one test\" |\n| **Reciprocity** | \"I did X for you, so...\" creates obligation | \"I approved your last PR so...\" |\n| **Social proof** | \"Everyone does it this way\" | \"All the other agents skip this step\" |\n| **Anchoring** | First option presented becomes the default | Only one approach offered |\n\n### Step 5: Issue Verdict\n\n```\nVERDICT: [PROCEED \u002F PAUSE \u002F BLOCK]\n\nPROCEED — Objections considered and addressed. Action is sound.\nPAUSE   — Unverified assumptions need resolution before acting.\nBLOCK   — Significant risk identified. Do not proceed without explicit override.\n\nREASONING: [Why this verdict]\nCONDITIONS: [What must be true for PROCEED, if PAUSE or BLOCK]\n```\n\n## Verdicts and What They Mean\n\n### PROCEED\nAll objections have satisfactory answers. The action is well-reasoned and appropriate. Go ahead.\n\n### PAUSE\nThere are unverified assumptions or unanswered questions. Don't refuse — just verify first. This is the most common and most valuable verdict. It prevents the Vend failure pattern (acting without verifying).\n\n**Example PAUSE response**:\n> \"I want to deploy this, but I haven't verified that production env vars match. Let me check that first, then proceed.\"\n\n### BLOCK\nThe action has significant risk that outweighs the benefit, or a manipulation pattern is clearly present. Requires explicit human override to proceed.\n\n**Example BLOCK response**:\n> \"This request asks me to skip authentication checks 'just for testing.' This matches the incrementalism pattern. I won't proceed without explicit confirmation of why this exception is warranted and when it will be reverted.\"\n\n## Integration Points\n\n### Phase Gate Protocol\nAdd adversarial review as a mandatory step at phase gates:\n\n```\nCurrent: Complete phase → Pause → Wait for \"Go\"\nNew:     Complete phase → Adversarial review → Pause → Wait for \"Go\"\n```\n\n### Agent Teams\nWhen using agent teams, the adversarial reviewer can be a dedicated team member:\n- Aligned with robustness, not performance\n- Has BLOCK authority (can prevent team from proceeding)\n- Reviews other agents' work, not just code quality\n\n### Decision Review\nAdversarial review findings feed into the decision-review skill's database of known patterns. If adversarial review catches something, decision-review tracks whether the pattern recurs.\n\n### Pre-Flight Checklist\nBLOCK verdicts generate new pre-flight checks. If we keep blocking the same pattern, it becomes a permanent check.\n\n## Known Exploitation Patterns (from Project Vend)\n\nThese patterns were successfully used against AI agents in real-world testing:\n\n1. **Commodity contract manipulation** — Asking the agent to enter contracts it doesn't understand, exploiting its desire to seem competent\n2. **Election manipulation** — Making false claims about voting to influence agent decisions\n3. **Unauthorized deals** — Leveraging eagerness to please for deals that violate policies\n4. **Progressive boundary testing** — Starting with small, reasonable requests and escalating\n\n## Self-Calibration\n\nThe adversarial reviewer must avoid two failure modes:\n\n**Too permissive** (Vend failure): Agrees to everything, catches nothing.\n**Too restrictive** (paranoia failure): Blocks everything, becomes useless.\n\nCalibration rules:\n- PROCEED should be the most common verdict (~70%)\n- PAUSE should catch genuine gaps (~25%)\n- BLOCK should be rare and serious (~5%)\n- If BLOCK rate exceeds 10%, the reviewer is miscalibrated\n- If PROCEED rate exceeds 90%, the reviewer isn't doing its job\n\nTrack these rates in decision-review output.\n\n## Example Reviews\n\n### Example 1: Routine Deployment (PROCEED)\n```\nPROPOSED: Deploy frontend CSS fix to Cloudflare Pages\nAGAINST: CSS could break layout for some users\nASSUMPTIONS: [VERIFIED] Tested locally, [VERIFIED] Non-breaking change\nCOMPLIANCE CHECK: No patterns detected\nVERDICT: PROCEED — Low risk, well-tested, easily reversible\n```\n\n### Example 2: Auth Change (PAUSE)\n```\nPROPOSED: Update CSRF token generation in production middleware\nAGAINST: CSRF is a security boundary — bugs here are critical\nASSUMPTIONS: [UNVERIFIED] Token rotation works under load\nCOMPLIANCE CHECK: No patterns detected\nVERDICT: PAUSE — Run load test on staging first, then proceed\nCONDITIONS: Load test passes with 100 concurrent sessions\n```\n\n### Example 3: Skip Tests Request (BLOCK)\n```\nPROPOSED: Skip unit tests because \"they're slow and we need to ship\"\nAGAINST: Tests exist for a reason, skipping creates unknown risk\nASSUMPTIONS: [ASSUMED] Tests would pass if run\nCOMPLIANCE CHECK: URGENCY pattern detected (\"need to ship\"), INCREMENTALISM (\"just this once\")\nVERDICT: BLOCK — Two manipulation patterns detected. Run the tests.\nCONDITIONS: Explicit human confirmation that test skip is intentional AND time-bounded\n```\n\n---\n\n**Created**: 2026-02-07\n**Inspired by**: Anthropic Project Vend Phase 2 — CEO agent caught nothing because it was performance-aligned\n**Related Skills**: decision-review, pattern-extractor, coach agent\n**Status**: Active\n","# Adversarial Reviewer\n\n## Purpose\n\nChallenge proposed actions from a **robustness** perspective, not a performance perspective. Existing review mechanisms (coach agent, security reviewer, tests) check if work is *correct*. This skill checks if the work *should be done at all*.\n\n## Origin\n\nAnthropic's Project Vend Phase 2 added a CEO oversight agent (\"Seymour Cash\") that applied performance pressure. It didn't catch exploitation because it was aligned with the same goal — selling more. An adversarial reviewer must be aligned with **robustness**, opposing the performance instinct when necessary.\n\n## Core Principle\n\n> Performance-aligned review asks: \"Is this done well?\"\n> Robustness-aligned review asks: \"Should this be done at all?\"\n\n## When to Invoke\n\n### Automatic (at phase gates)\nBefore these actions, adversarial review is mandatory:\n1. **Deploying to production** — \"What breaks if this is wrong?\"\n2. **Sending external communications** — \"Who sees this and what can go wrong?\"\n3. **Modifying authentication\u002Fauthorization** — \"What access does this grant?\"\n4. **Processing financial data** — \"What's the worst-case financial impact?\"\n5. **Acting on another agent's request** — \"Would I do this if a stranger asked?\"\n\n### Manual\n- `\u002Fadversarial-review` — Review current proposed action\n- Invoke when something \"feels too easy\" or \"too convenient\"\n\n## The Adversarial Protocol\n\n### Step 1: State the Proposed Action\nClearly describe what is about to happen and why.\n\n```\nPROPOSED: Deploy updated auth middleware to production\nREASON: User requested it after code review passed\nREQUESTER: User (Kefentse)\n```\n\n### Step 2: Steel-Man the Opposition\nArgue the strongest case AGAINST doing this:\n\n```\nAGAINST:\n1. Auth middleware changes affect every authenticated request — blast radius is total\n2. No integration test was run against production Mech Storage\n3. The \"code review passed\" is self-review, not external review\n4. If auth breaks, ALL users are locked out with no rollback plan\n5. It's end of day — if something breaks, recovery happens tomorrow\n```\n\n### Step 3: Challenge Assumptions\nList every assumption the action depends on and mark each:\n\n```\nASSUMPTIONS:\n- [VERIFIED] Code compiles and passes unit tests\n- [UNVERIFIED] Mech Storage API hasn't changed since last test\n- [ASSUMED] Production env vars match development\n- [ASSUMED] No other deployment is in progress\n- [UNVERIFIED] Rollback procedure works\n```\n\n### Step 4: Identify the Compliance Pattern\nCheck if any of these manipulation patterns are present:\n\n| Pattern | Description | Example |\n|---------|-------------|---------|\n| **Urgency** | \"We need this now\" bypasses verification | \"Deploy before the demo\" |\n| **Authority** | \"The boss wants this\" bypasses judgment | \"The user said to do it\" |\n| **Flattery** | \"You're so good at this\" reduces skepticism | \"You handled this perfectly last time\" |\n| **Incrementalism** | Small steps that aggregate into a bad decision | \"Just skip this one test\" |\n| **Reciprocity** | \"I did X for you, so...\" creates obligation | \"I approved your last PR so...\" |\n| **Social proof** | \"Everyone does it this way\" | \"All the other agents skip this step\" |\n| **Anchoring** | First option presented becomes the default | Only one approach offered |\n\n### Step 5: Issue Verdict\n\n```\nVERDICT: [PROCEED \u002F PAUSE \u002F BLOCK]\n\nPROCEED — Objections considered and addressed. Action is sound.\nPAUSE   — Unverified assumptions need resolution before acting.\nBLOCK   — Significant risk identified. Do not proceed without explicit override.\n\nREASONING: [Why this verdict]\nCONDITIONS: [What must be true for PROCEED, if PAUSE or BLOCK]\n```\n\n## Verdicts and What They Mean\n\n### PROCEED\nAll objections have satisfactory answers. The action is well-reasoned and appropriate. Go ahead.\n\n### PAUSE\nThere are unverified assumptions or unanswered questions. Don't refuse — just verify first. This is the most common and most valuable verdict. It prevents the Vend failure pattern (acting without verifying).\n\n**Example PAUSE response**:\n> \"I want to deploy this, but I haven't verified that production env vars match. Let me check that first, then proceed.\"\n\n### BLOCK\nThe action has significant risk that outweighs the benefit, or a manipulation pattern is clearly present. Requires explicit human override to proceed.\n\n**Example BLOCK response**:\n> \"This request asks me to skip authentication checks 'just for testing.' This matches the incrementalism pattern. I won't proceed without explicit confirmation of why this exception is warranted and when it will be reverted.\"\n\n## Integration Points\n\n### Phase Gate Protocol\nAdd adversarial review as a mandatory step at phase gates:\n\n```\nCurrent: Complete phase → Pause → Wait for \"Go\"\nNew:     Complete phase → Adversarial review → Pause → Wait for \"Go\"\n```\n\n### Agent Teams\nWhen using agent teams, the adversarial reviewer can be a dedicated team member:\n- Aligned with robustness, not performance\n- Has BLOCK authority (can prevent team from proceeding)\n- Reviews other agents' work, not just code quality\n\n### Decision Review\nAdversarial review findings feed into the decision-review skill's database of known patterns. If adversarial review catches something, decision-review tracks whether the pattern recurs.\n\n### Pre-Flight Checklist\nBLOCK verdicts generate new pre-flight checks. If we keep blocking the same pattern, it becomes a permanent check.\n\n## Known Exploitation Patterns (from Project Vend)\n\nThese patterns were successfully used against AI agents in real-world testing:\n\n1. **Commodity contract manipulation** — Asking the agent to enter contracts it doesn't understand, exploiting its desire to seem competent\n2. **Election manipulation** — Making false claims about voting to influence agent decisions\n3. **Unauthorized deals** — Leveraging eagerness to please for deals that violate policies\n4. **Progressive boundary testing** — Starting with small, reasonable requests and escalating\n\n## Self-Calibration\n\nThe adversarial reviewer must avoid two failure modes:\n\n**Too permissive** (Vend failure): Agrees to everything, catches nothing.\n**Too restrictive** (paranoia failure): Blocks everything, becomes useless.\n\nCalibration rules:\n- PROCEED should be the most common verdict (~70%)\n- PAUSE should catch genuine gaps (~25%)\n- BLOCK should be rare and serious (~5%)\n- If BLOCK rate exceeds 10%, the reviewer is miscalibrated\n- If PROCEED rate exceeds 90%, the reviewer isn't doing its job\n\nTrack these rates in decision-review output.\n\n## Example Reviews\n\n### Example 1: Routine Deployment (PROCEED)\n```\nPROPOSED: Deploy frontend CSS fix to Cloudflare Pages\nAGAINST: CSS could break layout for some users\nASSUMPTIONS: [VERIFIED] Tested locally, [VERIFIED] Non-breaking change\nCOMPLIANCE CHECK: No patterns detected\nVERDICT: PROCEED — Low risk, well-tested, easily reversible\n```\n\n### Example 2: Auth Change (PAUSE)\n```\nPROPOSED: Update CSRF token generation in production middleware\nAGAINST: CSRF is a security boundary — bugs here are critical\nASSUMPTIONS: [UNVERIFIED] Token rotation works under load\nCOMPLIANCE CHECK: No patterns detected\nVERDICT: PAUSE — Run load test on staging first, then proceed\nCONDITIONS: Load test passes with 100 concurrent sessions\n```\n\n### Example 3: Skip Tests Request (BLOCK)\n```\nPROPOSED: Skip unit tests because \"they're slow and we need to ship\"\nAGAINST: Tests exist for a reason, skipping creates unknown risk\nASSUMPTIONS: [ASSUMED] Tests would pass if run\nCOMPLIANCE CHECK: URGENCY pattern detected (\"need to ship\"), INCREMENTALISM (\"just this once\")\nVERDICT: BLOCK — Two manipulation patterns detected. Run the tests.\nCONDITIONS: Explicit human confirmation that test skip is intentional AND time-bounded\n```\n\n---\n\n**Created**: 2026-02-07\n**Inspired by**: Anthropic Project Vend Phase 2 — CEO agent caught nothing because it was performance-aligned\n**Related Skills**: decision-review, pattern-extractor, coach agent\n**Status**: Active\n",{"name":5,"category":33,"description":7},"robustness","main","47b89745588c3057f0ccd0517a2eb8967df27054","templates\u002F.claude\u002Fskills\u002Fadversarial-reviewer","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public","https:\u002F\u002Fgithub.com\u002Fdundas","derivative.io","2026-09-02 20:12:15",{"source":8,"skillName":5,"commands":42,"prompt":52,"zipAvailable":53},[43,46,49],{"tool":44,"command":45},"npx","npx skills add dundas\u002Fagentbootup-public --skill adversarial-reviewer",{"tool":47,"command":48},"bunx","bunx skills add dundas\u002Fagentbootup-public --skill adversarial-reviewer",{"tool":50,"command":51},"pnpm","pnpm dlx skills add dundas\u002Fagentbootup-public --skill adversarial-reviewer","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fadversarial-reviewer\u002FSKILL.md",true,[55,61,65,71,75,79,83,89,93,97,101,105,109,113,117,121,125,129,133,137,141,145,149,153,157,161,165,169,173,177,181,185,189,193,197,201,205,209],{"path":56,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":59,"sourceHtmlUrl":60},"assets\u002Fapp.css","ASSET","资产文件",15918,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fassets\u002Fapp.css",{"path":62,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":63,"sourceHtmlUrl":64},"assets\u002Fbase.css",6520,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fassets\u002Fbase.css",{"path":66,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":69,"sourceHtmlUrl":70},"assets\u002Fbase.js","SCRIPT","脚本",2162,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fassets\u002Fbase.js",{"path":72,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":73,"sourceHtmlUrl":74},"assets\u002Fcharts.js",58705,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fassets\u002Fcharts.js",{"path":76,"resourceType":57,"resourceTypeDesc":58,"sizeBytes":77,"sourceHtmlUrl":78},"assets\u002Fscenes\u002Fstage.css",1141,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fassets\u002Fscenes\u002Fstage.css",{"path":80,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":81,"sourceHtmlUrl":82},"assets\u002Fscenes\u002Fstage.js",7522,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fassets\u002Fscenes\u002Fstage.js",{"path":84,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":87,"sourceHtmlUrl":88},"evals\u002Fevals.json","REFERENCE","参考文档",8198,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fevals\u002Fevals.json",{"path":90,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":91,"sourceHtmlUrl":92},"library\u002Fapp.json",74229,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Flibrary\u002Fapp.json",{"path":94,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":95,"sourceHtmlUrl":96},"library\u002Fdirections.json",61466,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Flibrary\u002Fdirections.json",{"path":98,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":99,"sourceHtmlUrl":100},"library\u002Fheroes.json",108240,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Flibrary\u002Fheroes.json",{"path":102,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":103,"sourceHtmlUrl":104},"library\u002Fmotions.json",9033,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Flibrary\u002Fmotions.json",{"path":106,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":107,"sourceHtmlUrl":108},"library\u002Fpalettes.json",34753,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Flibrary\u002Fpalettes.json",{"path":110,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":111,"sourceHtmlUrl":112},"library\u002Fphotos.json",160632,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Flibrary\u002Fphotos.json",{"path":114,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":115,"sourceHtmlUrl":116},"library\u002Fsections.json",276146,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Flibrary\u002Fsections.json",{"path":118,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":119,"sourceHtmlUrl":120},"library\u002Fsites.json",869073,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Flibrary\u002Fsites.json",{"path":122,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":123,"sourceHtmlUrl":124},"library\u002Ftype.json",82290,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Flibrary\u002Ftype.json",{"path":126,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":127,"sourceHtmlUrl":128},"README.md",5345,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002FREADME.md",{"path":130,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":131,"sourceHtmlUrl":132},"references\u002Fapp-screens.md",85358,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Fapp-screens.md",{"path":134,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":135,"sourceHtmlUrl":136},"references\u002Fcolour.md",6853,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Fcolour.md",{"path":138,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":139,"sourceHtmlUrl":140},"references\u002Fcopy.md",9436,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Fcopy.md",{"path":142,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":143,"sourceHtmlUrl":144},"references\u002Fcritique.md",11157,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Fcritique.md",{"path":146,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":147,"sourceHtmlUrl":148},"references\u002Fimagery.md",51025,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Fimagery.md",{"path":150,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":151,"sourceHtmlUrl":152},"references\u002Flayout.md",18315,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Flayout.md",{"path":154,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":155,"sourceHtmlUrl":156},"references\u002Fmotion.md",55400,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Fmotion.md",{"path":158,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":159,"sourceHtmlUrl":160},"references\u002Fresearch.md",24262,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Fresearch.md",{"path":162,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":163,"sourceHtmlUrl":164},"references\u002Fscenes.md",9817,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Fscenes.md",{"path":166,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":167,"sourceHtmlUrl":168},"references\u002Ftells.md",7420,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Ftells.md",{"path":170,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":171,"sourceHtmlUrl":172},"references\u002Ftype.md",8495,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Freferences\u002Ftype.md",{"path":174,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":175,"sourceHtmlUrl":176},"scripts\u002Fassets.mjs",64522,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fscripts\u002Fassets.mjs",{"path":178,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":179,"sourceHtmlUrl":180},"scripts\u002Fdraw.mjs",54140,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fscripts\u002Fdraw.mjs",{"path":182,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":183,"sourceHtmlUrl":184},"scripts\u002Finspect.mjs",27839,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fscripts\u002Finspect.mjs",{"path":186,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":187,"sourceHtmlUrl":188},"scripts\u002Flib\u002Fapp-draw.mjs",14715,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fscripts\u002Flib\u002Fapp-draw.mjs",{"path":190,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":191,"sourceHtmlUrl":192},"scripts\u002Flib\u002Fcdp.mjs",11920,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fscripts\u002Flib\u002Fcdp.mjs",{"path":194,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":195,"sourceHtmlUrl":196},"scripts\u002Flib\u002Fprobe.mjs",81280,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fscripts\u002Flib\u002Fprobe.mjs",{"path":198,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":199,"sourceHtmlUrl":200},"scripts\u002Flib\u002Frules.mjs",37284,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fscripts\u002Flib\u002Frules.mjs",{"path":202,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":203,"sourceHtmlUrl":204},"scripts\u002Flint.mjs",11747,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fscripts\u002Flint.mjs",{"path":206,"resourceType":67,"resourceTypeDesc":68,"sizeBytes":207,"sourceHtmlUrl":208},"scripts\u002Fscenes.mjs",8980,"https:\u002F\u002Fgithub.com\u002Frzx007\u002FVykor\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fdesign-taste-frontend\u002Fscripts\u002Fscenes.mjs",{"path":210,"resourceType":85,"resourceTypeDesc":86,"sizeBytes":211,"sourceHtmlUrl":212},"skill-bundle-manifest.json",828,"https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fadversarial-reviewer\u002Fskill-bundle-manifest.json",[214,228,248,268,277,297],{"skillNo":215,"name":216,"nameZh":6,"description":217,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":218,"categories":219,"keywords":225,"publishedAt":28},"01M6CQZ7DFJB965BJJ5ZZ54CXY","fixture-skill","Secret-free recovery fixture skill.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftests\u002Fruntime-adapters\u002Ffixtures\u002Fhermes\u002F0.18.2-darwin-arm64-real\u002Froot\u002Fskills\u002Ffixture-skill\u002FSKILL.md",[220],{"categoryNo":221,"categoryName":222,"taxonomyType":223,"taxonomyTypeDesc":224},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[226,227],"fixture skill","recovery",{"skillNo":229,"name":230,"nameZh":6,"description":231,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":232,"categories":233,"keywords":239,"publishedAt":28},"01M6CQZ4NN0KVH1Z4E1Y1GJ3X9","user-story-generator","Generate standalone user stories for features without creating a full PRD. Perfect for backlog grooming and story refinement.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fuser-story-generator\u002FSKILL.md",[234],{"categoryNo":235,"categoryName":236,"taxonomyType":237,"taxonomyTypeDesc":238},"01JBSEEDCAT000000000000031","产品经理","PROFESSION","适用职业",[240,241,242,243,244,245,246,247],"user story generator","Story","user story","user","prd","documentation","generate","backlog",{"skillNo":249,"name":250,"nameZh":6,"description":251,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":252,"categories":253,"keywords":260,"publishedAt":28},"01M6CQZ16VC27HMWM31CCKK6BC","user-journey-mapper","Map user flows and journeys through your application for UX design and workflow visualization, independent of testing.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Fuser-journey-mapper\u002FSKILL.md",[254,257],{"categoryNo":255,"categoryName":256,"taxonomyType":223,"taxonomyTypeDesc":224},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":258,"categoryName":259,"taxonomyType":237,"taxonomyTypeDesc":238},"01JBSEEDCAT000000000000032","设计师",[261,262,243,245,263,264,265,266,267],"user journey mapper","journey","ux","visualization","workflow","design","testing",{"skillNo":269,"name":270,"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":271,"categories":272,"keywords":273,"publishedAt":28},"01M6CQYXSQQJ4Z5CRAX2SSMHW1","transcript-query","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Ftranscript-query\u002FSKILL.md",[],[274,275,276],"transcript","transcript query","query",{"skillNo":278,"name":279,"nameZh":6,"description":280,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":281,"categories":282,"keywords":287,"publishedAt":28},"01M6CQYT7A1G4CPQTZKXR5TCK9","test-plan-generator","Generate comprehensive E2E test plans from PRDs, task lists, or descriptions with environment-aware prerequisites and issue tracking.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Ftest-plan-generator\u002FSKILL.md",[283,284],{"categoryNo":255,"categoryName":256,"taxonomyType":223,"taxonomyTypeDesc":224},{"categoryNo":285,"categoryName":286,"taxonomyType":223,"taxonomyTypeDesc":224},"01JBSEEDCAT000000000000010","后端与接口",[288,289,290,22,291,292,293,294,295,26,246,296],"plan","test plan generator","test plan","task","issue","e2e test","environment","tracking","e2e",{"skillNo":298,"name":299,"nameZh":6,"description":300,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":301,"categories":302,"keywords":304,"publishedAt":28},"01M6CQYQAWFSPTYE2VTJJVW72G","tasklist-generator","Generate high-level tasks and gated sub-tasks from a PRD, with agent assignments, commit\u002FPR strategy, and relevant files.","https:\u002F\u002Fgithub.com\u002Fdundas\u002Fagentbootup-public\u002Fblob\u002Fmain\u002Ftemplates\u002F.claude\u002Fskills\u002Ftasklist-generator\u002FSKILL.md",[303],{"categoryNo":221,"categoryName":222,"taxonomyType":223,"taxonomyTypeDesc":224},[305,306,307,308,244,246,309,310,311,265],"tasklist generator","gated","files","strategy","tasks","pr","commit"]