[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7BFMYHRQAY2NEJM5CH83G2P":3,"skill-same:01M7BFMYHRQAY2NEJM5CH83G2P":49},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":31,"ownerHtmlUrl":32,"ownerBlogUrl":6,"ownerDescription":33,"updatedAt":34,"install":35,"resources":48},"01M7BFMYHRQAY2NEJM5CH83G2P","agent-architecture-audit",null,"Full-stack diagnostic for agent and LLM applications. Audits the 12-layer agent stack for wrapper regression, memory pollution, tool discipline failures, hidden repair loops, and rendering corruption. Produces severity-ranked findings with code-first fixes. Essential for developers building agent applications, autonomous loops, or any LLM-powered feature.","YYC-Cube\u002FYYC3-Knowledge-Base","YYC-Cube","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F235006543?v=4","01M79J2PH4SWD2X0825X6PN748","yyc3","01M79J2PHVK3F6ZBYT8K2SCZCW",1,"MIT","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-AI-应用工具\u002FYYC3-AI-应用工具-Everything-claude-code\u002Fskills\u002Fagent-architecture-audit\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: agent-architecture-audit\ndescription: Full-stack diagnostic for agent and LLM applications. Audits the 12-layer agent stack for wrapper regression, memory pollution, tool discipline failures, hidden repair loops, and rendering corruption. Produces severity-ranked findings with code-first fixes. Essential for developers building agent applications, autonomous loops, or any LLM-powered feature.\norigin: oh-my-agent-check\ntools: Read, Write, Edit, Bash, Grep, Glob\n---\n\n# Agent Architecture Audit\n\nA diagnostic workflow for agent systems that hide failures behind wrapper layers, stale memory, retry loops, or transport\u002Frendering mutations.\n\n## When to Activate\n\n**MANDATORY for:**\n- Releasing any agent or LLM-powered application to production\n- Shipping features with tool calling, memory, or multi-step workflows\n- Agent behavior degrades after adding wrapper layers\n- User reports \"the agent is getting worse\" or \"tools are flaky\"\n- Same model works in playground but breaks inside your wrapper\n- Debugging agent behavior for more than 15 minutes without finding root cause\n\n**Especially critical when:**\n- You've added new prompt layers, tool definitions, or memory systems\n- Different agents in your system behave inconsistently\n- The model was fine yesterday but is hallucinating today\n- You suspect hidden repair\u002Fretry loops silently mutating responses\n\n**Do not use for:**\n- General code debugging — use `agent-introspection-debugging`\n- Code review — use language-specific reviewer agents\n- Security scanning — use `security-review` or `security-review\u002Fscan`\n- Agent performance benchmarking — use `agent-eval`\n- Writing new features — use the appropriate workflow skill\n\n## The 12-Layer Stack\n\nEvery agent system has these layers. Any of them can corrupt the answer:\n\n| # | Layer | What Goes Wrong |\n|---|-------|----------------|\n| 1 | System prompt | Conflicting instructions, instruction bloat |\n| 2 | Session history | Stale context injection from previous turns |\n| 3 | Long-term memory | Pollution across sessions, old topics in new conversations |\n| 4 | Distillation | Compressed artifacts re-entering as pseudo-facts |\n| 5 | Active recall | Redundant re-summary layers wasting context |\n| 6 | Tool selection | Wrong tool routing, model skips required tools |\n| 7 | Tool execution | Hallucinated execution — claims to call but doesn't |\n| 8 | Tool interpretation | Misread or ignored tool output |\n| 9 | Answer shaping | Format corruption in final response |\n| 10 | Platform rendering | Transport-layer mutation (UI, API, CLI mutates valid answers) |\n| 11 | Hidden repair loops | Silent fallback\u002Fretry agents running second LLM pass |\n| 12 | Persistence | Expired state or cached artifacts reused as live evidence |\n\n## Common Failure Patterns\n\n### 1. Wrapper Regression\n\nThe base model produces correct answers, but the wrapper layers make it worse.\n\n**Symptoms:**\n- Model works fine in playground or direct API call, breaks in your agent\n- Added a new prompt layer, existing behavior degraded\n- Agent sounds confident but is confidently wrong\n- \"It was working before the last update\"\n\n### 2. Memory Contamination\n\nOld topics leak into new conversations through history, memory retrieval, or distillation.\n\n**Symptoms:**\n- Agent brings up unrelated past topics\n- User corrections don't stick (old memory overwrites new)\n- Same-session artifacts re-enter as pseudo-facts\n- Memory grows without bound, degrading response quality over time\n\n### 3. Tool Discipline Failure\n\nTools are declared in the prompt but not enforced in code. The model skips them or hallucinates execution.\n\n**Symptoms:**\n- \"Must use tool X\" in prompt, but model answers without calling it\n- Tool results look correct but were never actually executed\n- Different tools fight over the same responsibility\n- Model uses tool when it shouldn't, or skips it when it must\n\n### 4. Rendering\u002FTransport Corruption\n\nThe agent's internal answer is correct, but the platform layer mutates it during delivery.\n\n**Symptoms:**\n- Logs show correct answer, user sees broken output\n- Markdown rendering, JSON parsing, or streaming fragments corrupt valid responses\n- Hidden fallback agent quietly replaces the answer before delivery\n- Output differs between terminal and UI\n\n### 5. Hidden Agent Layers\n\nSilent repair, retry, summarization, or recall agents run without explicit contracts.\n\n**Symptoms:**\n- Output changes between internal generation and user delivery\n- \"Auto-fix\" loops run a second LLM pass the user doesn't know about\n- Multiple agents modify the same output without coordination\n- Answers get \"smoothed\" or \"corrected\" by invisible layers\n\n## Audit Workflow\n\n### Phase 1: Scope\n\nDefine what you're auditing:\n\n- **Target system** — what agent application?\n- **Entrypoints** — how do users interact with it?\n- **Model stack** — which LLM(s) and providers?\n- **Symptoms** — what does the user report?\n- **Time window** — when did it start?\n- **Layers to audit** — which of the 12 layers apply?\n\n### Phase 2: Evidence Collection\n\nGather evidence from the codebase:\n\n- **Source code** — agent loop, tool router, memory admission, prompt assembly\n- **Logs** — historical session traces, tool call records\n- **Config** — prompt templates, tool schemas, provider settings\n- **Memory files** — SOPs, knowledge bases, session archives\n\nUse `rg` to search for anti-patterns:\n\n```bash\n# Tool requirements expressed only in prompt text (not code)\nrg \"must.*tool|必须.*工具|required.*call\" --type md\n\n# Tool execution without validation\nrg \"tool_call|toolCall|tool_use\" --type py --type ts\n\n# Hidden LLM calls outside main agent loop\nrg \"completion|chat\\.create|messages\\.create|llm\\.invoke\"\n\n# Memory admission without user-correction priority\nrg \"memory.*admit|long.*term.*update|persist.*memory\" --type py --type ts\n\n# Fallback loops that run additional LLM calls\nrg \"fallback|retry.*llm|repair.*prompt|re-?prompt\" --type py --type ts\n\n# Silent output mutation\nrg \"mutate|rewrite.*response|transform.*output|shap\" --type py --type ts\n```\n\n### Phase 3: Failure Mapping\n\nFor each finding, document:\n\n- **Symptom** — what the user sees\n- **Mechanism** — how the wrapper causes it\n- **Source layer** — which of the 12 layers\n- **Root cause** — the deepest cause\n- **Evidence** — file:line or log:row reference\n- **Confidence** — 0.0 to 1.0\n\n### Phase 4: Fix Strategy\n\nDefault fix order (code-first, not prompt-first):\n\n1. **Code-gate tool requirements** — enforce in code, not just prompt text\n2. **Remove or narrow hidden repair agents** — make fallback explicit with contracts\n3. **Reduce context duplication** — same info through prompt + history + memory + distillation\n4. **Tighten memory admission** — user corrections > agent assertions\n5. **Tighten distillation triggers** — don't compress what shouldn't be compressed\n6. **Reduce rendering mutation** — pass-through, don't transform\n7. **Convert to typed JSON envelopes** — structured internal flow, not freeform prose\n\n## Severity Model\n\n| Level | Meaning | Action |\n|-------|---------|--------|\n| `critical` | Agent can confidently produce wrong operational behavior | Fix before next release |\n| `high` | Agent frequently degrades correctness or stability | Fix this sprint |\n| `medium` | Correctness usually survives but output is fragile or wasteful | Plan for next cycle |\n| `low` | Mostly cosmetic or maintainability issues | Backlog |\n\n## Output Format\n\nPresent findings to the user in this order:\n\n1. **Severity-ranked findings** (most critical first)\n2. **Architecture diagnosis** (which layer corrupted what, and why)\n3. **Ordered fix plan** (code-first, not prompt-first)\n\nDo not lead with compliments or summaries. If the system is broken, say so directly.\n\n## Quick Diagnostic Questions\n\nWhen auditing an agent system, answer these:\n\n| # | Question | If Yes → |\n|---|----------|----------|\n| 1 | Can the model skip a required tool and still answer? | Tool not code-gated |\n| 2 | Does old conversation content appear in new turns? | Memory contamination |\n| 3 | Is the same info in system prompt AND memory AND history? | Context duplication |\n| 4 | Does the platform run a second LLM pass before delivery? | Hidden repair loop |\n| 5 | Does the output differ between internal generation and user delivery? | Rendering corruption |\n| 6 | Are \"must use tool X\" rules only in prompt text? | Tool discipline failure |\n| 7 | Can the agent's own monologue become persistent memory? | Memory poisoning |\n\n## Anti-Patterns to Avoid\n\n- Avoid blaming the model before falsifying wrapper-layer regressions.\n- Avoid blaming memory without showing the contamination path.\n- Do not let a clean current state erase a dirty historical incident.\n- Do not treat markdown prose as a trustworthy internal protocol.\n- Do not accept \"must use tool\" in prompt text when code never enforces it.\n- Keep findings direct, evidence-backed, and severity-ranked.\n\n## Report Schema\n\nAudits should produce structured reports following this shape:\n\n```json\n{\n  \"schema_version\": \"ecc.agent-architecture-audit.report.v1\",\n  \"executive_verdict\": {\n    \"overall_health\": \"high_risk\",\n    \"primary_failure_mode\": \"string\",\n    \"most_urgent_fix\": \"string\"\n  },\n  \"scope\": {\n    \"target_name\": \"string\",\n    \"model_stack\": [\"string\"],\n    \"layers_to_audit\": [\"string\"]\n  },\n  \"findings\": [\n    {\n      \"severity\": \"critical|high|medium|low\",\n      \"title\": \"string\",\n      \"mechanism\": \"string\",\n      \"source_layer\": \"string\",\n      \"root_cause\": \"string\",\n      \"evidence_refs\": [\"file:line\"],\n      \"confidence\": 0.0,\n      \"recommended_fix\": \"string\"\n    }\n  ],\n  \"ordered_fix_plan\": [\n    { \"order\": 1, \"goal\": \"string\", \"why_now\": \"string\", \"expected_effect\": \"string\" }\n  ]\n}\n```\n\n## Related Skills\n\n- `agent-introspection-debugging` — Debug agent runtime failures (loops, timeouts, state errors)\n- `agent-eval` — Benchmark agent performance head-to-head\n- `security-review` — Security audit for code and configuration\n- `autonomous-agent-harness` — Set up autonomous agent operations\n- `agent-harness-construction` — Build agent harnesses from scratch\n","# Agent Architecture Audit\n\nA diagnostic workflow for agent systems that hide failures behind wrapper layers, stale memory, retry loops, or transport\u002Frendering mutations.\n\n## When to Activate\n\n**MANDATORY for:**\n- Releasing any agent or LLM-powered application to production\n- Shipping features with tool calling, memory, or multi-step workflows\n- Agent behavior degrades after adding wrapper layers\n- User reports \"the agent is getting worse\" or \"tools are flaky\"\n- Same model works in playground but breaks inside your wrapper\n- Debugging agent behavior for more than 15 minutes without finding root cause\n\n**Especially critical when:**\n- You've added new prompt layers, tool definitions, or memory systems\n- Different agents in your system behave inconsistently\n- The model was fine yesterday but is hallucinating today\n- You suspect hidden repair\u002Fretry loops silently mutating responses\n\n**Do not use for:**\n- General code debugging — use `agent-introspection-debugging`\n- Code review — use language-specific reviewer agents\n- Security scanning — use `security-review` or `security-review\u002Fscan`\n- Agent performance benchmarking — use `agent-eval`\n- Writing new features — use the appropriate workflow skill\n\n## The 12-Layer Stack\n\nEvery agent system has these layers. Any of them can corrupt the answer:\n\n| # | Layer | What Goes Wrong |\n|---|-------|----------------|\n| 1 | System prompt | Conflicting instructions, instruction bloat |\n| 2 | Session history | Stale context injection from previous turns |\n| 3 | Long-term memory | Pollution across sessions, old topics in new conversations |\n| 4 | Distillation | Compressed artifacts re-entering as pseudo-facts |\n| 5 | Active recall | Redundant re-summary layers wasting context |\n| 6 | Tool selection | Wrong tool routing, model skips required tools |\n| 7 | Tool execution | Hallucinated execution — claims to call but doesn't |\n| 8 | Tool interpretation | Misread or ignored tool output |\n| 9 | Answer shaping | Format corruption in final response |\n| 10 | Platform rendering | Transport-layer mutation (UI, API, CLI mutates valid answers) |\n| 11 | Hidden repair loops | Silent fallback\u002Fretry agents running second LLM pass |\n| 12 | Persistence | Expired state or cached artifacts reused as live evidence |\n\n## Common Failure Patterns\n\n### 1. Wrapper Regression\n\nThe base model produces correct answers, but the wrapper layers make it worse.\n\n**Symptoms:**\n- Model works fine in playground or direct API call, breaks in your agent\n- Added a new prompt layer, existing behavior degraded\n- Agent sounds confident but is confidently wrong\n- \"It was working before the last update\"\n\n### 2. Memory Contamination\n\nOld topics leak into new conversations through history, memory retrieval, or distillation.\n\n**Symptoms:**\n- Agent brings up unrelated past topics\n- User corrections don't stick (old memory overwrites new)\n- Same-session artifacts re-enter as pseudo-facts\n- Memory grows without bound, degrading response quality over time\n\n### 3. Tool Discipline Failure\n\nTools are declared in the prompt but not enforced in code. The model skips them or hallucinates execution.\n\n**Symptoms:**\n- \"Must use tool X\" in prompt, but model answers without calling it\n- Tool results look correct but were never actually executed\n- Different tools fight over the same responsibility\n- Model uses tool when it shouldn't, or skips it when it must\n\n### 4. Rendering\u002FTransport Corruption\n\nThe agent's internal answer is correct, but the platform layer mutates it during delivery.\n\n**Symptoms:**\n- Logs show correct answer, user sees broken output\n- Markdown rendering, JSON parsing, or streaming fragments corrupt valid responses\n- Hidden fallback agent quietly replaces the answer before delivery\n- Output differs between terminal and UI\n\n### 5. Hidden Agent Layers\n\nSilent repair, retry, summarization, or recall agents run without explicit contracts.\n\n**Symptoms:**\n- Output changes between internal generation and user delivery\n- \"Auto-fix\" loops run a second LLM pass the user doesn't know about\n- Multiple agents modify the same output without coordination\n- Answers get \"smoothed\" or \"corrected\" by invisible layers\n\n## Audit Workflow\n\n### Phase 1: Scope\n\nDefine what you're auditing:\n\n- **Target system** — what agent application?\n- **Entrypoints** — how do users interact with it?\n- **Model stack** — which LLM(s) and providers?\n- **Symptoms** — what does the user report?\n- **Time window** — when did it start?\n- **Layers to audit** — which of the 12 layers apply?\n\n### Phase 2: Evidence Collection\n\nGather evidence from the codebase:\n\n- **Source code** — agent loop, tool router, memory admission, prompt assembly\n- **Logs** — historical session traces, tool call records\n- **Config** — prompt templates, tool schemas, provider settings\n- **Memory files** — SOPs, knowledge bases, session archives\n\nUse `rg` to search for anti-patterns:\n\n```bash\n# Tool requirements expressed only in prompt text (not code)\nrg \"must.*tool|必须.*工具|required.*call\" --type md\n\n# Tool execution without validation\nrg \"tool_call|toolCall|tool_use\" --type py --type ts\n\n# Hidden LLM calls outside main agent loop\nrg \"completion|chat\\.create|messages\\.create|llm\\.invoke\"\n\n# Memory admission without user-correction priority\nrg \"memory.*admit|long.*term.*update|persist.*memory\" --type py --type ts\n\n# Fallback loops that run additional LLM calls\nrg \"fallback|retry.*llm|repair.*prompt|re-?prompt\" --type py --type ts\n\n# Silent output mutation\nrg \"mutate|rewrite.*response|transform.*output|shap\" --type py --type ts\n```\n\n### Phase 3: Failure Mapping\n\nFor each finding, document:\n\n- **Symptom** — what the user sees\n- **Mechanism** — how the wrapper causes it\n- **Source layer** — which of the 12 layers\n- **Root cause** — the deepest cause\n- **Evidence** — file:line or log:row reference\n- **Confidence** — 0.0 to 1.0\n\n### Phase 4: Fix Strategy\n\nDefault fix order (code-first, not prompt-first):\n\n1. **Code-gate tool requirements** — enforce in code, not just prompt text\n2. **Remove or narrow hidden repair agents** — make fallback explicit with contracts\n3. **Reduce context duplication** — same info through prompt + history + memory + distillation\n4. **Tighten memory admission** — user corrections > agent assertions\n5. **Tighten distillation triggers** — don't compress what shouldn't be compressed\n6. **Reduce rendering mutation** — pass-through, don't transform\n7. **Convert to typed JSON envelopes** — structured internal flow, not freeform prose\n\n## Severity Model\n\n| Level | Meaning | Action |\n|-------|---------|--------|\n| `critical` | Agent can confidently produce wrong operational behavior | Fix before next release |\n| `high` | Agent frequently degrades correctness or stability | Fix this sprint |\n| `medium` | Correctness usually survives but output is fragile or wasteful | Plan for next cycle |\n| `low` | Mostly cosmetic or maintainability issues | Backlog |\n\n## Output Format\n\nPresent findings to the user in this order:\n\n1. **Severity-ranked findings** (most critical first)\n2. **Architecture diagnosis** (which layer corrupted what, and why)\n3. **Ordered fix plan** (code-first, not prompt-first)\n\nDo not lead with compliments or summaries. If the system is broken, say so directly.\n\n## Quick Diagnostic Questions\n\nWhen auditing an agent system, answer these:\n\n| # | Question | If Yes → |\n|---|----------|----------|\n| 1 | Can the model skip a required tool and still answer? | Tool not code-gated |\n| 2 | Does old conversation content appear in new turns? | Memory contamination |\n| 3 | Is the same info in system prompt AND memory AND history? | Context duplication |\n| 4 | Does the platform run a second LLM pass before delivery? | Hidden repair loop |\n| 5 | Does the output differ between internal generation and user delivery? | Rendering corruption |\n| 6 | Are \"must use tool X\" rules only in prompt text? | Tool discipline failure |\n| 7 | Can the agent's own monologue become persistent memory? | Memory poisoning |\n\n## Anti-Patterns to Avoid\n\n- Avoid blaming the model before falsifying wrapper-layer regressions.\n- Avoid blaming memory without showing the contamination path.\n- Do not let a clean current state erase a dirty historical incident.\n- Do not treat markdown prose as a trustworthy internal protocol.\n- Do not accept \"must use tool\" in prompt text when code never enforces it.\n- Keep findings direct, evidence-backed, and severity-ranked.\n\n## Report Schema\n\nAudits should produce structured reports following this shape:\n\n```json\n{\n  \"schema_version\": \"ecc.agent-architecture-audit.report.v1\",\n  \"executive_verdict\": {\n    \"overall_health\": \"high_risk\",\n    \"primary_failure_mode\": \"string\",\n    \"most_urgent_fix\": \"string\"\n  },\n  \"scope\": {\n    \"target_name\": \"string\",\n    \"model_stack\": [\"string\"],\n    \"layers_to_audit\": [\"string\"]\n  },\n  \"findings\": [\n    {\n      \"severity\": \"critical|high|medium|low\",\n      \"title\": \"string\",\n      \"mechanism\": \"string\",\n      \"source_layer\": \"string\",\n      \"root_cause\": \"string\",\n      \"evidence_refs\": [\"file:line\"],\n      \"confidence\": 0.0,\n      \"recommended_fix\": \"string\"\n    }\n  ],\n  \"ordered_fix_plan\": [\n    { \"order\": 1, \"goal\": \"string\", \"why_now\": \"string\", \"expected_effect\": \"string\" }\n  ]\n}\n```\n\n## Related Skills\n\n- `agent-introspection-debugging` — Debug agent runtime failures (loops, timeouts, state errors)\n- `agent-eval` — Benchmark agent performance head-to-head\n- `security-review` — Security audit for code and configuration\n- `autonomous-agent-harness` — Set up autonomous agent operations\n- `agent-harness-construction` — Build agent harnesses from scratch\n",{"name":5,"tools":24,"origin":25,"description":7},"Read, Write, Edit, Bash, Grep, Glob","oh-my-agent-check","main","87afe6ce96220accf50a345dea46784c8c9f210e","YYC3-AI-应用工具\u002FYYC3-AI-应用工具-Everything-claude-code\u002Fskills\u002Fagent-architecture-audit",0,"https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base","https:\u002F\u002Fkb.yyc3.top\u002F","https:\u002F\u002Fgithub.com\u002FYYC-Cube","万象归元于云枢 | 深栈智启新纪元","2026-11-10 10:37:32",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":47},[37,40,43],{"tool":38,"command":39},"npx","npx skills add YYC-Cube\u002FYYC3-Knowledge-Base --skill agent-architecture-audit",{"tool":41,"command":42},"bunx","bunx skills add YYC-Cube\u002FYYC3-Knowledge-Base --skill agent-architecture-audit",{"tool":44,"command":45},"pnpm","pnpm dlx skills add YYC-Cube\u002FYYC3-Knowledge-Base --skill agent-architecture-audit","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-AI-应用工具\u002FYYC3-AI-应用工具-Everything-claude-code\u002Fskills\u002Fagent-architecture-audit\u002FSKILL.md",true,[],[50,57,64,71,78,85],{"skillNo":51,"name":52,"nameZh":6,"description":53,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":19},"01M7CTDWHERCXPQCP0K03A86AS","vercel-sandbox","Run agent-browser + Chrome inside Vercel Sandbox microVMs for browser automation from any Vercel-deployed app. Use when the user needs browser automation in a Vercel app (Next.js, SvelteKit, Nuxt, Remix, Astro, etc.), wants to run headless Chrome without binary size limits, needs persistent browser sessions across commands, or wants ephemeral isolated browser environments. Triggers include \"Vercel Sandbox browser\", \"microVM Chrome\", \"agent-browser in sandbox\", \"browser automation on Vercel\", or any task requiring Chrome in a Vercel Sandbox.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Fvercel-sandbox\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M7CTD8D2BNJY54S65M47HEJA","slack","Interact with Slack workspaces using browser automation. Use when the user needs to check unread channels, navigate Slack, send messages, extract data, find information, search conversations, or automate any Slack task. Triggers include \"check my Slack\", \"what channels have unreads\", \"send a message to\", \"search Slack for\", \"extract from Slack\", \"find who said\", or any task requiring programmatic Slack interaction.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Fslack\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M7CTCSSB8ZP3Q58C3YZDZK1W","electron","Automate Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify, etc.) using agent-browser via Chrome DevTools Protocol. Use when the user needs to interact with an Electron app, automate a desktop app, connect to a running app, control a native app, or test an Electron application. Triggers include \"automate Slack app\", \"control VS Code\", \"interact with Discord app\", \"test this Electron app\", \"connect to desktop app\", or any task requiring automation of a native Electron application.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Felectron\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7CTCBJY76688V4DCZGWFHGQ","dogfood","Systematically explore and test a web application to find bugs, UX issues, and other problems. Use when asked to \"dogfood\", \"QA\", \"exploratory test\", \"find issues\", \"bug hunt\", \"test this app\u002Fsite\u002Fplatform\", or review the quality of a web application. Produces a structured report with full reproduction evidence -- step-by-step screenshots, repro videos, and detailed repro steps for every issue -- so findings can be handed directly to the responsible teams.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Fdogfood\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":19},"01M7CTC4H0FFGVEAQR9PA4GA1A","agent-browser","Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to \"open a website\", \"fill out a form\", \"click a button\", \"take a screenshot\", \"scrape data from a page\", \"test this web app\", \"login to a site\", \"automate browser actions\", or any task requiring programmatic web interaction.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-WebSpec-Web标准\u002FYYC3-WebSpec-Web标准-standards\u002Fbrowser-agent\u002Fskills\u002Fagent-browser\u002FSKILL.md",[],[],{"skillNo":86,"name":87,"nameZh":6,"description":88,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":89,"categories":90,"keywords":91,"publishedAt":19},"01M7CTBKVFYBG78FDSABSDJTTQ","shadcn","Manages shadcn components and projects — adding, searching, fixing, debugging, styling, and composing UI, including chat interfaces. Provides project context, component docs, and usage examples. Applies when working with shadcn\u002Fui, component registries, presets, --preset codes, or any project with a components.json file. Also triggers for \"shadcn init\", \"create an app with --preset\", \"switch to --preset\", and for building a design system or applying a DESIGN.md or brand spec to shadcn\u002Fui.","https:\u002F\u002Fgithub.com\u002FYYC-Cube\u002FYYC3-Knowledge-Base\u002Fblob\u002Fmain\u002FYYC3-UI-界面组件\u002FYYC3-UI-界面组件-ui\u002Fskills\u002Fshadcn\u002FSKILL.md",[],[]]