{
  "schemaVersion": "1.0",
  "item": {
    "slug": "afrexai-ai-coding-toolkit",
    "name": "AI Coding Toolkit",
    "source": "tencent",
    "type": "skill",
    "category": "开发工具",
    "sourceUrl": "https://clawhub.ai/1kalin/afrexai-ai-coding-toolkit",
    "canonicalUrl": "https://clawhub.ai/1kalin/afrexai-ai-coding-toolkit",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadMode": "redirect",
    "downloadUrl": "/downloads/afrexai-ai-coding-toolkit",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-ai-coding-toolkit",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "installMethod": "Manual import",
    "extraction": "Extract archive",
    "prerequisites": [
      "OpenClaw"
    ],
    "packageFormat": "ZIP package",
    "includedAssets": [
      "README.md",
      "SKILL.md"
    ],
    "primaryDoc": "SKILL.md",
    "quickSetup": [
      "Download the package from Yavira.",
      "Extract the archive and review SKILL.md first.",
      "Import or place the package into your OpenClaw setup."
    ],
    "agentAssist": {
      "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
      "steps": [
        "Download the package from Yavira.",
        "Extract it into a folder your agent can access.",
        "Paste one of the prompts below and point your agent at the extracted folder."
      ],
      "prompts": [
        {
          "label": "New install",
          "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete."
        },
        {
          "label": "Upgrade existing",
          "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run."
        }
      ]
    },
    "sourceHealth": {
      "source": "tencent",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-04-23T16:43:11.935Z",
      "expiresAt": "2026-04-30T16:43:11.935Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=4claw-imageboard",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=4claw-imageboard",
        "contentDisposition": "attachment; filename=\"4claw-imageboard-1.0.1.zip\"",
        "redirectLocation": null,
        "bodySnippet": null
      },
      "scope": "source",
      "summary": "Source download looks usable.",
      "detail": "Yavira can redirect you to the upstream package for this source.",
      "primaryActionLabel": "Download for OpenClaw",
      "primaryActionHref": "/downloads/afrexai-ai-coding-toolkit"
    },
    "validation": {
      "installChecklist": [
        "Use the Yavira download entry.",
        "Review SKILL.md after the package is downloaded.",
        "Confirm the extracted package contains the expected setup assets."
      ],
      "postInstallChecks": [
        "Confirm the extracted package includes the expected docs or setup files.",
        "Validate the skill or prompts are available in your target agent workspace.",
        "Capture any manual follow-up steps the agent could not complete."
      ]
    },
    "downloadPageUrl": "https://openagent3.xyz/downloads/afrexai-ai-coding-toolkit",
    "agentPageUrl": "https://openagent3.xyz/skills/afrexai-ai-coding-toolkit/agent",
    "manifestUrl": "https://openagent3.xyz/skills/afrexai-ai-coding-toolkit/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/afrexai-ai-coding-toolkit/agent.md"
  },
  "agentAssist": {
    "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
    "steps": [
      "Download the package from Yavira.",
      "Extract it into a folder your agent can access.",
      "Paste one of the prompts below and point your agent at the extracted folder."
    ],
    "prompts": [
      {
        "label": "New install",
        "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete."
      },
      {
        "label": "Upgrade existing",
        "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run."
      }
    ]
  },
  "documentation": {
    "source": "clawhub",
    "primaryDoc": "SKILL.md",
    "sections": [
      {
        "title": "AI Coding Toolkit — Master Every AI Coding Assistant",
        "body": "The complete methodology for 10X productivity with AI-assisted development. Covers Cursor, Windsurf, Cline, Aider, Claude Code, GitHub Copilot, and more — tool-agnostic principles that work everywhere."
      },
      {
        "title": "Phase 1: Quick Assessment — Where Are You?",
        "body": "Rate yourself 1-5 on each:\n\nDimension1 (Beginner)5 (Expert)Prompt quality\"Fix this bug\"Structured context + constraints + examplesContext managementPaste entire filesCurated context windows, .cursorrules, AGENTS.mdWorkflow integrationAd-hoc usageSystematic agent-first developmentOutput verificationAccept everythingReview, test, iterate before committingTool selectionOne tool for everythingRight tool for right task\n\nScore interpretation:\n\n5-10: Read everything — you'll 10X your output\n11-18: Skip to Phase 4+ for advanced techniques\n19-25: Focus on Phase 8-10 for mastery patterns"
      },
      {
        "title": "Decision Guide: Which AI Coding Tool When?",
        "body": "ToolBest ForContext WindowAutonomy LevelCostGitHub CopilotLine/function completion, inline suggestionsCurrent file + neighborsLow (autocomplete)$10-19/moCursorFull-file editing, multi-file refactors, chatProject-aware (indexing)Medium (tab/chat/composer)$20/moWindsurf (Cascade)Autonomous multi-step tasks, flowsProject-aware + flowsHigh (agentic flows)$15/moClineVS Code extension, model-agnostic, transparentManual context + autoHigh (tool use, browser)API costsAiderTerminal-based, git-native, pair programmingRepo map + selected filesMedium-High (git commits)API costsClaude CodeCLI agent, complex multi-file tasksWorkspace-awareHigh (full agent)API costsOpenClawPersistent agent, cron, multi-surfaceWorkspace + memory + toolsVery High (autonomous)API costs"
      },
      {
        "title": "Selection Decision Tree",
        "body": "Need autocomplete while typing?\n  → GitHub Copilot (layer it with any other tool)\n\nWorking in VS Code/IDE?\n  ├─ Want integrated editor experience? → Cursor or Windsurf\n  ├─ Want model flexibility + transparency? → Cline\n  └─ Want minimal config, just works? → Cursor\n\nWorking in terminal?\n  ├─ Want git-native pair programming? → Aider\n  ├─ Want full agent with tools? → Claude Code\n  └─ Want persistent autonomous agent? → OpenClaw\n\nBuilding complex multi-file features?\n  → Cursor Composer or Windsurf Cascade or Claude Code\n\nNeed autonomous background work?\n  → OpenClaw (cron, heartbeats, multi-session)"
      },
      {
        "title": "Recommended Stack (Layer These)",
        "body": "Solo developer:\n\nGitHub Copilot (always-on autocomplete)\nCursor OR Windsurf (primary IDE)\nClaude Code OR Aider (terminal agent for complex tasks)\n\nTeam:\n\nGitHub Copilot (org-wide)\nCursor (primary IDE, .cursorrules in repo)\nCI/CD AI review (automated PR review)"
      },
      {
        "title": "Phase 3: Context Engineering — The #1 Skill",
        "body": "Context is everything. The quality of AI output is directly proportional to the quality of context you provide."
      },
      {
        "title": "The Context Hierarchy (Most → Least Important)",
        "body": "System instructions (.cursorrules, AGENTS.md, CLAUDE.md, .windsurfrules)\nExplicit context (files you @mention or add to chat)\nImplicit context (open tabs, recent edits, project index)\nModel knowledge (training data — least reliable for your codebase)"
      },
      {
        "title": "Project Rules File Template",
        "body": "Create at project root. Name depends on tool:\n\nCursor: .cursorrules\nWindsurf: .windsurfrules\nClaude Code: CLAUDE.md\nAider: .aider.conf.yml + convention docs\nOpenClaw: AGENTS.md\n\n# [PROJECT] — AI Coding Context\n\n## Project Overview\n- Name: [project name]\n- Stack: [e.g., Next.js 14 + TypeScript + Tailwind + Drizzle + PostgreSQL]\n- Architecture: [e.g., App Router, server components by default]\n- Monorepo: [yes/no, structure if yes]\n\n## Code Standards (ENFORCE STRICTLY)\n- TypeScript strict mode (`tsc --noEmit --strict`)\n- Max 50 lines per function, 300 lines per file\n- One responsibility per file\n- Naming: camelCase functions, PascalCase types, SCREAMING_SNAKE constants\n- Imports: named imports, no default exports\n- Error handling: explicit try/catch, typed errors, no silent catches\n\n## Patterns to Follow\n- [Pattern 1 with example]\n- [Pattern 2 with example]\n- [Pattern 3 with example]\n\n## Anti-Patterns (NEVER DO)\n- [Anti-pattern 1]\n- [Anti-pattern 2]\n- [Anti-pattern 3]\n\n## File Structure\n\nsrc/\ncomponents/     # React components\nlib/            # Shared utilities\nserver/         # Server-only code\ndb/             # Database schema + queries\ntypes/          # Shared TypeScript types\n\n## Testing\n- Framework: [vitest/jest/pytest]\n- Pattern: AAA (Arrange, Act, Assert)\n- Naming: `should [expected behavior] when [condition]`\n- Coverage target: [80%+]\n\n## Dependencies\n- Approved: [list]\n- Banned: [list with reasons]\n\n## Common Commands\n- `npm run dev` — start dev server\n- `npm run test` — run tests\n- `npm run lint` — lint + typecheck\n- `npm run build` — production build"
      },
      {
        "title": "Context Window Management",
        "body": "The 80/20 Rule: 80% of your context should be the specific files/functions relevant to the task. 20% is project conventions and standards.\n\nContext Compression Techniques:\n\nSummarize, don't dump — Instead of pasting a 500-line file, describe what it does and paste only the relevant section\nUse @mentions — @file.ts instead of copy-paste (tool-specific)\nCreate reference docs — One-page architecture summaries the AI can reference\nPrune conversation — Start new chats for new tasks; stale context = hallucinations\nTree command — Give the AI your project structure: tree -I node_modules -L 3"
      },
      {
        "title": "The Context Refresh Rule",
        "body": "Every 5-10 messages, check: Is the AI still tracking correctly?\nIf it starts hallucinating file names, functions, or making wrong assumptions — start a new chat with fresh context.\nContext is milk. It spoils."
      },
      {
        "title": "The SPEC Framework (Structure, Precision, Examples, Constraints)",
        "body": "Bad prompt:\n\nFix the login bug\n\nGood prompt (SPEC):\n\n## Structure\nFix the authentication flow in `src/auth/login.ts`\n\n## Precision\n- The login function throws \"user not found\" even when the user exists\n- Error occurs on line 42 when querying by email (case-sensitive match)\n- PostgreSQL query uses exact match but emails are stored lowercase\n\n## Examples\n- Input: \"User@Example.com\" → should match \"user@example.com\" in DB\n- Current behavior: returns null\n- Expected: returns user record\n\n## Constraints\n- Don't change the database schema\n- Use the existing `normalizeEmail()` utility from `src/utils/email.ts`\n- Add a test case for case-insensitive lookup\n- Keep the existing error handling pattern (throw AppError)"
      },
      {
        "title": "Prompt Templates by Task Type",
        "body": "Feature Implementation:\n\nImplement [feature] in [file/location].\n\nRequirements:\n1. [Requirement with acceptance criteria]\n2. [Requirement with acceptance criteria]\n3. [Requirement with acceptance criteria]\n\nConstraints:\n- Follow existing patterns in [reference file]\n- Use [specific library/approach]\n- Include error handling for [edge cases]\n- Write tests in [test file location]\n\nReference: Here's how similar feature [X] was implemented:\n[paste relevant code snippet]\n\nBug Fix:\n\nBug: [description]\nFile: [path]\nSteps to reproduce: [1, 2, 3]\nExpected: [behavior]\nActual: [behavior]\nError: [paste error message/stack trace]\n\nFix constraints:\n- Don't change [protected areas]\n- Add regression test\n- Explain root cause before fixing\n\nRefactoring:\n\nRefactor [file/module] to [goal].\n\nCurrent state: [describe current architecture]\nTarget state: [describe desired architecture]\nMotivation: [why — performance, readability, maintainability]\n\nRules:\n- Preserve all existing behavior (no functional changes)\n- Keep all existing tests passing\n- Break into small, reviewable commits\n- Each commit should be independently deployable\n\nCode Review:\n\nReview this code for:\n1. Correctness — logic errors, edge cases, race conditions\n2. Security — injection, auth bypass, data exposure\n3. Performance — N+1 queries, unnecessary allocations, missing indexes\n4. Maintainability — naming, complexity, test coverage\n\nBe specific: quote the line, explain the issue, suggest the fix.\nSkip style/formatting — linter handles that.\n\n[paste code]"
      },
      {
        "title": "Pattern 1: Test-Driven AI Development (TDD-AI)",
        "body": "1. Write the test first (yourself or with AI help)\n2. Ask AI to implement the code that passes the test\n3. Run tests — verify green\n4. Ask AI to refactor while keeping tests green\n5. Review the final code yourself\n\nWhy this works: Tests are specifications. The AI writes better code when it has a concrete target. You catch hallucinations immediately."
      },
      {
        "title": "Pattern 2: Scaffold → Fill → Review",
        "body": "1. Ask AI to scaffold the architecture (file structure, interfaces, types)\n2. Review and approve the scaffold\n3. Ask AI to fill in implementation file by file\n4. Review each file individually\n5. Integration test the full feature\n\nWhy this works: You maintain architectural control. The AI handles the grunt work. Errors are caught at each layer."
      },
      {
        "title": "Pattern 3: Conversation Threading",
        "body": "Chat 1: Architecture discussion → decisions documented\nChat 2: Implementation of Component A (reference architecture doc)\nChat 3: Implementation of Component B (reference architecture doc)\nChat 4: Integration + testing\n\nWhy this works: Fresh context per component prevents drift. Architecture doc provides continuity."
      },
      {
        "title": "Pattern 4: AI Pair Programming (Aider/Claude Code)",
        "body": "1. Start session with repo context\n2. Describe the task in natural language\n3. AI proposes changes as git diffs\n4. Review each diff before accepting\n5. AI commits with meaningful messages\n6. You handle edge cases and integration"
      },
      {
        "title": "Pattern 5: Autonomous Agent Workflow (OpenClaw/Claude Code)",
        "body": "1. Define task in structured format (acceptance criteria, constraints)\n2. Agent plans → executes → verifies (reads files, runs tests)\n3. Agent creates PR/branch with changes\n4. You review the complete changeset\n5. Iterate on feedback"
      },
      {
        "title": "Cursor",
        "body": "FeaturePower MoveTab completionLet it complete 3-5 tokens before accepting — catches wrong predictions earlyCmd+K (inline edit)Select ONLY the exact lines to change — less context = more accurateChat@file to add context, @codebase for project-wide questionsComposerMulti-file changes — describe the full feature, let it edit across files.cursorrulesProject-specific AI instructions — commit to repo for team alignmentNotepadsReusable context (API docs, design docs) — attach to any chat\n\nCursor Pro Tips:\n\nUse @git to reference recent changes\nUse @docs to reference official library documentation\nCreate .cursor/rules/ directory for multiple rule files by domain\n\"Apply\" button to accept chat suggestions directly into code"
      },
      {
        "title": "Windsurf (Cascade)",
        "body": "FeaturePower MoveCascade flowsMulti-step autonomous tasks — it can read, write, run terminalWrite modeDirect file editing with AIChat modeDiscussion without editing.windsurfrulesProject context fileTurbo modeFaster, less accurate — good for simple tasks\n\nWindsurf Pro Tips:\n\nCascade excels at multi-file refactors — give it the full scope\nUse \"undo flow\" to revert entire multi-step changes\nPin important files in context\nLet it read error output from terminal to self-fix"
      },
      {
        "title": "Cline",
        "body": "FeaturePower MoveModel selectionSwitch models per task (cheap for simple, expensive for complex)Tool useReads files, runs commands, opens browser — full agentTransparencyShows every action before executing — audit everythingCustom instructionsPer-project system promptsAuto-approveConfigure which actions need approval\n\nCline Pro Tips:\n\nSet spending limits to prevent runaway API costs\nUse cheaper models (Haiku/GPT-4o-mini) for simple tasks\nEnable \"diff mode\" to see exact changes before applying\nCreate task-specific instruction files"
      },
      {
        "title": "Aider",
        "body": "FeaturePower Move/add filesExplicitly control which files the AI can see/edit/read filesRead-only context (reference files)/architectTwo-model approach — architect plans, editor implementsRepo mapAuto-generates codebase summary for contextGit integrationEvery change is a commit — easy rollback\n\nAider Pro Tips:\n\nUse --architect flag for complex features (planner + implementer)\n/drop files you don't need to free context window\n--map-tokens to control repo map size\nRun aider --model claude-sonnet-4-20250514 for best code quality"
      },
      {
        "title": "Claude Code",
        "body": "FeaturePower MoveFull agentReads files, writes code, runs tests, git operationsCLAUDE.mdProject instructions file — auto-loadedSub-agentsSpawn parallel workers for complex tasksMemoryPersistent across sessions (project-level)\n\nClaude Code Pro Tips:\n\nWrite a comprehensive CLAUDE.md — it's your biggest leverage\nUse \"plan mode\" first for complex tasks, then \"implement\"\nLet it run tests and self-correct — don't interrupt the loop\nUse /compact when context gets long"
      },
      {
        "title": "The Trust-But-Verify Checklist",
        "body": "After every AI-generated change:\n\nRead every line — don't blindly accept. AI hallucinates plausible-looking code\n Check imports — AI often imports non-existent modules or wrong versions\n Verify function signatures — parameter names, types, return types\n Test edge cases — AI optimizes for the happy path\n Check for security — hardcoded secrets, missing auth checks, SQL injection\n Run the tests — if tests pass, good. If no tests exist, write them first\n Check for drift — did it change files you didn't ask it to change?\n Verify dependencies — did it add packages? Are they real? Are they secure?"
      },
      {
        "title": "Common AI Code Failures",
        "body": "FailureDetectionFixHallucinated APICode uses functions that don't existCheck library docs before acceptingOutdated patternsUses deprecated APIs (React class components)Specify versions in contextMissing error handlingHappy path only, no try/catchAsk specifically for error casesSecurity holesInline secrets, missing auth, XSSSecurity review as separate stepOver-engineering5 files for a 20-line solutionAsk for simplest possible solutionWrong abstractionsPremature generalizationSpecify \"don't abstract, keep concrete\"Test theaterTests that pass but test nothingReview test assertions specificallyCopy-paste bugsDuplicated logic with subtle differencesCheck for patterns, extract helpers"
      },
      {
        "title": "The 3-Read Review",
        "body": "Skim read — Does the structure make sense? Right files, right approach?\nLogic read — Does each function do what it claims? Edge cases handled?\nIntegration read — Does it work with the rest of the codebase? Breaking changes?"
      },
      {
        "title": "Token Cost Awareness",
        "body": "ModelInput $/1M tokensOutput $/1M tokensBest ForGPT-4o mini$0.15$0.60Simple completions, formattingClaude Haiku$0.25$1.25Quick edits, simple questionsGPT-4o$2.50$10.00Complex code generationClaude Sonnet$3.00$15.00Complex code, long contextClaude Opus$15.00$75.00Architecture, hardest problemso3$10.00$40.00Complex reasoning, algorithms"
      },
      {
        "title": "Cost Reduction Strategies",
        "body": "Tier your usage — Simple tasks → cheap model. Complex → expensive model\nReduce context — Every unnecessary file in context costs money\nStart new chats — Long conversations accumulate expensive history\nUse autocomplete for simple stuff — Copilot is flat-rate, much cheaper per completion\nCache project context — Use rules files instead of re-explaining every chat\nBatch related tasks — Handle related changes in one conversation"
      },
      {
        "title": "Monthly Cost Benchmarks (Full-Time Developer)",
        "body": "Usage LevelEstimated Monthly CostLight (Copilot + occasional chat)$20-40Medium (Cursor Pro + daily chat)$40-80Heavy (API-based agents, complex tasks)$80-200Power user (autonomous agents, all day)$200-500+"
      },
      {
        "title": "Rolling Out AI Coding Tools to a Team",
        "body": "Week 1-2: Foundation\n\nChoose primary tool (Cursor or Windsurf recommended for teams)\nCreate .cursorrules / .windsurfrules committed to repo\nRun a 1-hour workshop: basics, prompt techniques, verification\nSet team guidelines (review requirements, security rules)\n\nWeek 3-4: Practice\n\nDaily 15-min \"AI wins\" standup share\nPair sessions: experienced + new user\nCollect common prompts into team prompt library\nMonitor and address concerns (quality, dependency)\n\nMonth 2: Optimization\n\nMeasure: time-to-PR, bugs-per-feature, developer satisfaction\nIterate on .cursorrules based on team feedback\nCreate task-specific prompt templates in shared docs\nAddress skill gaps: who's using it well, who needs help?\n\nMonth 3: Systemization\n\nAI-assisted PR review as CI step\nAutomated test generation for new features\nCustom slash commands / snippets for team workflows\nQuarterly review: ROI, quality metrics, tooling updates"
      },
      {
        "title": "Team Guidelines Template",
        "body": "# AI Coding Guidelines — [Team Name]\n\n## Approved Tools\n- [Tool 1] for [use case]\n- [Tool 2] for [use case]\n\n## Rules\n1. AI-generated code gets the SAME review rigor as human code\n2. Never paste proprietary/customer data into AI tools without approved data handling\n3. All AI-generated tests must be reviewed for assertion quality\n4. Security-sensitive code (auth, payments, PII) requires human-first approach\n5. Commit messages should NOT mention AI — own the code you commit\n\n## Quality Gates\n- [ ] Typecheck passes (`tsc --noEmit --strict`)\n- [ ] All tests pass\n- [ ] No new warnings\n- [ ] Manual review of all AI-generated code\n- [ ] Security-sensitive areas reviewed by security champion"
      },
      {
        "title": "Multi-Agent Architecture for Development",
        "body": "Task: Build feature X\n\nAgent 1 (Architect): Plans the approach, defines interfaces\nAgent 2 (Implementer): Writes the code\nAgent 3 (Tester): Writes and runs tests\nAgent 4 (Reviewer): Reviews for quality, security, patterns\n\nOrchestrator: Coordinates, resolves conflicts, maintains context"
      },
      {
        "title": "Self-Healing Development Loop",
        "body": "1. Agent writes code\n2. Agent runs tests\n3. Tests fail → agent reads error, fixes code\n4. Repeat until tests pass\n5. Agent runs linter\n6. Lint fails → agent fixes\n7. All green → create PR"
      },
      {
        "title": "The Prompt Library Pattern",
        "body": "Maintain a prompts/ directory in your project:\n\nprompts/\n  feature-implementation.md\n  bug-fix.md\n  refactoring.md\n  code-review.md\n  test-generation.md\n  migration.md\n  documentation.md\n\nEach file is a reusable prompt template. Reference them: \"Follow the template in prompts/feature-implementation.md\""
      },
      {
        "title": "Model Routing Strategy",
        "body": "task_routing:\n  autocomplete: copilot  # Always-on, flat rate\n  simple_edit: haiku     # Quick, cheap\n  feature_impl: sonnet   # Good balance\n  architecture: opus     # When it matters\n  debugging: sonnet      # Needs to reason about code\n  documentation: haiku   # Simple transformation\n  security_review: opus  # Can't afford mistakes\n  test_generation: sonnet # Needs understanding of code logic"
      },
      {
        "title": "Phase 11: Anti-Patterns — What NOT to Do",
        "body": "Anti-PatternWhy It FailsDo This InsteadPrompt and prayNo verification = bugs in productionAlways review, always testPaste the whole codebaseOverwhelms context, increases costCurate relevant files onlyNever start new chatsStale context → hallucinationsNew task = new chatTrust without readingAI generates plausible but wrong codeRead every lineSkip tests because AI wrote itAI code has bugs tooTest AI code MORE, not lessUse one model for everythingWaste money on simple tasksTier models by complexityNo project rules fileAI guesses your conventionsWrite .cursorrules / CLAUDE.mdVague promptsGarbage in, garbage outUse SPEC frameworkOver-relianceSkill atrophy, can't debug AI outputUnderstand what AI generatesIgnoring securityAI doesn't prioritize securityExplicit security review step"
      },
      {
        "title": "AI-Assisted Development Quality Score (0-100)",
        "body": "DimensionWeightCriteriaContext engineering20%Rules files, curated context, fresh chatsPrompt quality15%SPEC framework, task-appropriate templatesVerification rigor20%Review checklist, test coverage, security reviewTool selection10%Right tool for task, model routingCost efficiency10%Tiered usage, context management, batch tasksOutput quality15%Code correctness, maintainability, no driftWorkflow integration10%Systematic process, team alignment"
      },
      {
        "title": "Weekly Self-Review Questions",
        "body": "What was my best AI-assisted output this week? What made it good?\nWhere did AI waste my time? What went wrong with context/prompts?\nAm I reviewing thoroughly enough, or rubber-stamping?\nWhat prompt patterns worked well? Add to prompt library.\nAm I over-relying on AI for things I should understand deeply?"
      },
      {
        "title": "Monthly Metrics",
        "body": "Acceleration factor: Tasks completed per day vs pre-AI baseline\nBug rate: Bugs in AI-assisted code vs manual code\nCost per feature: API spend / features shipped\nContext efficiency: Average conversation length before drift\nCoverage: % of codebase with AI-assisted tests"
      },
      {
        "title": "Quick Reference: Natural Language Commands",
        "body": "\"Set up AI coding for [project]\" — Generate rules file + tool recommendations\n\"Write a prompt for [task type]\" — Generate SPEC-formatted prompt template\n\"Review this AI output\" — Run the Trust-But-Verify checklist\n\"Compare [tool A] vs [tool B] for [use case]\" — Tool selection analysis\n\"Optimize my AI coding costs\" — Analyze usage and suggest model routing\n\"Create a team AI coding guide\" — Generate team guidelines document\n\"Debug why AI keeps [hallucinating X]\" — Context diagnosis\n\"Set up test-driven AI workflow for [feature]\" — TDD-AI pattern guide\n\"Create prompt library for [project type]\" — Generate prompt templates\n\"Score my AI coding maturity\" — Run the quality assessment\n\"Onboard [person] to AI coding\" — Generate training plan\n\"Audit AI coding security practices\" — Security review checklist"
      }
    ],
    "body": "AI Coding Toolkit — Master Every AI Coding Assistant\n\nThe complete methodology for 10X productivity with AI-assisted development. Covers Cursor, Windsurf, Cline, Aider, Claude Code, GitHub Copilot, and more — tool-agnostic principles that work everywhere.\n\nPhase 1: Quick Assessment — Where Are You?\n\nRate yourself 1-5 on each:\n\nDimension\t1 (Beginner)\t5 (Expert)\nPrompt quality\t\"Fix this bug\"\tStructured context + constraints + examples\nContext management\tPaste entire files\tCurated context windows, .cursorrules, AGENTS.md\nWorkflow integration\tAd-hoc usage\tSystematic agent-first development\nOutput verification\tAccept everything\tReview, test, iterate before committing\nTool selection\tOne tool for everything\tRight tool for right task\n\nScore interpretation:\n\n5-10: Read everything — you'll 10X your output\n11-18: Skip to Phase 4+ for advanced techniques\n19-25: Focus on Phase 8-10 for mastery patterns\nPhase 2: Tool Selection Matrix\nDecision Guide: Which AI Coding Tool When?\nTool\tBest For\tContext Window\tAutonomy Level\tCost\nGitHub Copilot\tLine/function completion, inline suggestions\tCurrent file + neighbors\tLow (autocomplete)\t$10-19/mo\nCursor\tFull-file editing, multi-file refactors, chat\tProject-aware (indexing)\tMedium (tab/chat/composer)\t$20/mo\nWindsurf (Cascade)\tAutonomous multi-step tasks, flows\tProject-aware + flows\tHigh (agentic flows)\t$15/mo\nCline\tVS Code extension, model-agnostic, transparent\tManual context + auto\tHigh (tool use, browser)\tAPI costs\nAider\tTerminal-based, git-native, pair programming\tRepo map + selected files\tMedium-High (git commits)\tAPI costs\nClaude Code\tCLI agent, complex multi-file tasks\tWorkspace-aware\tHigh (full agent)\tAPI costs\nOpenClaw\tPersistent agent, cron, multi-surface\tWorkspace + memory + tools\tVery High (autonomous)\tAPI costs\nSelection Decision Tree\nNeed autocomplete while typing?\n  → GitHub Copilot (layer it with any other tool)\n\nWorking in VS Code/IDE?\n  ├─ Want integrated editor experience? → Cursor or Windsurf\n  ├─ Want model flexibility + transparency? → Cline\n  └─ Want minimal config, just works? → Cursor\n\nWorking in terminal?\n  ├─ Want git-native pair programming? → Aider\n  ├─ Want full agent with tools? → Claude Code\n  └─ Want persistent autonomous agent? → OpenClaw\n\nBuilding complex multi-file features?\n  → Cursor Composer or Windsurf Cascade or Claude Code\n\nNeed autonomous background work?\n  → OpenClaw (cron, heartbeats, multi-session)\n\nRecommended Stack (Layer These)\n\nSolo developer:\n\nGitHub Copilot (always-on autocomplete)\nCursor OR Windsurf (primary IDE)\nClaude Code OR Aider (terminal agent for complex tasks)\n\nTeam:\n\nGitHub Copilot (org-wide)\nCursor (primary IDE, .cursorrules in repo)\nCI/CD AI review (automated PR review)\nPhase 3: Context Engineering — The #1 Skill\n\nContext is everything. The quality of AI output is directly proportional to the quality of context you provide.\n\nThe Context Hierarchy (Most → Least Important)\nSystem instructions (.cursorrules, AGENTS.md, CLAUDE.md, .windsurfrules)\nExplicit context (files you @mention or add to chat)\nImplicit context (open tabs, recent edits, project index)\nModel knowledge (training data — least reliable for your codebase)\nProject Rules File Template\n\nCreate at project root. Name depends on tool:\n\nCursor: .cursorrules\nWindsurf: .windsurfrules\nClaude Code: CLAUDE.md\nAider: .aider.conf.yml + convention docs\nOpenClaw: AGENTS.md\n# [PROJECT] — AI Coding Context\n\n## Project Overview\n- Name: [project name]\n- Stack: [e.g., Next.js 14 + TypeScript + Tailwind + Drizzle + PostgreSQL]\n- Architecture: [e.g., App Router, server components by default]\n- Monorepo: [yes/no, structure if yes]\n\n## Code Standards (ENFORCE STRICTLY)\n- TypeScript strict mode (`tsc --noEmit --strict`)\n- Max 50 lines per function, 300 lines per file\n- One responsibility per file\n- Naming: camelCase functions, PascalCase types, SCREAMING_SNAKE constants\n- Imports: named imports, no default exports\n- Error handling: explicit try/catch, typed errors, no silent catches\n\n## Patterns to Follow\n- [Pattern 1 with example]\n- [Pattern 2 with example]\n- [Pattern 3 with example]\n\n## Anti-Patterns (NEVER DO)\n- [Anti-pattern 1]\n- [Anti-pattern 2]\n- [Anti-pattern 3]\n\n## File Structure\n\n\nsrc/ components/ # React components lib/ # Shared utilities server/ # Server-only code db/ # Database schema + queries types/ # Shared TypeScript types\n\n\n## Testing\n- Framework: [vitest/jest/pytest]\n- Pattern: AAA (Arrange, Act, Assert)\n- Naming: `should [expected behavior] when [condition]`\n- Coverage target: [80%+]\n\n## Dependencies\n- Approved: [list]\n- Banned: [list with reasons]\n\n## Common Commands\n- `npm run dev` — start dev server\n- `npm run test` — run tests\n- `npm run lint` — lint + typecheck\n- `npm run build` — production build\n\nContext Window Management\n\nThe 80/20 Rule: 80% of your context should be the specific files/functions relevant to the task. 20% is project conventions and standards.\n\nContext Compression Techniques:\n\nSummarize, don't dump — Instead of pasting a 500-line file, describe what it does and paste only the relevant section\nUse @mentions — @file.ts instead of copy-paste (tool-specific)\nCreate reference docs — One-page architecture summaries the AI can reference\nPrune conversation — Start new chats for new tasks; stale context = hallucinations\nTree command — Give the AI your project structure: tree -I node_modules -L 3\nThe Context Refresh Rule\n\nEvery 5-10 messages, check: Is the AI still tracking correctly? If it starts hallucinating file names, functions, or making wrong assumptions — start a new chat with fresh context. Context is milk. It spoils.\n\nPhase 4: Prompt Engineering for Code\nThe SPEC Framework (Structure, Precision, Examples, Constraints)\n\nBad prompt:\n\nFix the login bug\n\n\nGood prompt (SPEC):\n\n## Structure\nFix the authentication flow in `src/auth/login.ts`\n\n## Precision\n- The login function throws \"user not found\" even when the user exists\n- Error occurs on line 42 when querying by email (case-sensitive match)\n- PostgreSQL query uses exact match but emails are stored lowercase\n\n## Examples\n- Input: \"User@Example.com\" → should match \"user@example.com\" in DB\n- Current behavior: returns null\n- Expected: returns user record\n\n## Constraints\n- Don't change the database schema\n- Use the existing `normalizeEmail()` utility from `src/utils/email.ts`\n- Add a test case for case-insensitive lookup\n- Keep the existing error handling pattern (throw AppError)\n\nPrompt Templates by Task Type\n\nFeature Implementation:\n\nImplement [feature] in [file/location].\n\nRequirements:\n1. [Requirement with acceptance criteria]\n2. [Requirement with acceptance criteria]\n3. [Requirement with acceptance criteria]\n\nConstraints:\n- Follow existing patterns in [reference file]\n- Use [specific library/approach]\n- Include error handling for [edge cases]\n- Write tests in [test file location]\n\nReference: Here's how similar feature [X] was implemented:\n[paste relevant code snippet]\n\n\nBug Fix:\n\nBug: [description]\nFile: [path]\nSteps to reproduce: [1, 2, 3]\nExpected: [behavior]\nActual: [behavior]\nError: [paste error message/stack trace]\n\nFix constraints:\n- Don't change [protected areas]\n- Add regression test\n- Explain root cause before fixing\n\n\nRefactoring:\n\nRefactor [file/module] to [goal].\n\nCurrent state: [describe current architecture]\nTarget state: [describe desired architecture]\nMotivation: [why — performance, readability, maintainability]\n\nRules:\n- Preserve all existing behavior (no functional changes)\n- Keep all existing tests passing\n- Break into small, reviewable commits\n- Each commit should be independently deployable\n\n\nCode Review:\n\nReview this code for:\n1. Correctness — logic errors, edge cases, race conditions\n2. Security — injection, auth bypass, data exposure\n3. Performance — N+1 queries, unnecessary allocations, missing indexes\n4. Maintainability — naming, complexity, test coverage\n\nBe specific: quote the line, explain the issue, suggest the fix.\nSkip style/formatting — linter handles that.\n\n[paste code]\n\nPhase 5: Workflow Patterns — Agent-First Development\nPattern 1: Test-Driven AI Development (TDD-AI)\n1. Write the test first (yourself or with AI help)\n2. Ask AI to implement the code that passes the test\n3. Run tests — verify green\n4. Ask AI to refactor while keeping tests green\n5. Review the final code yourself\n\n\nWhy this works: Tests are specifications. The AI writes better code when it has a concrete target. You catch hallucinations immediately.\n\nPattern 2: Scaffold → Fill → Review\n1. Ask AI to scaffold the architecture (file structure, interfaces, types)\n2. Review and approve the scaffold\n3. Ask AI to fill in implementation file by file\n4. Review each file individually\n5. Integration test the full feature\n\n\nWhy this works: You maintain architectural control. The AI handles the grunt work. Errors are caught at each layer.\n\nPattern 3: Conversation Threading\nChat 1: Architecture discussion → decisions documented\nChat 2: Implementation of Component A (reference architecture doc)\nChat 3: Implementation of Component B (reference architecture doc)\nChat 4: Integration + testing\n\n\nWhy this works: Fresh context per component prevents drift. Architecture doc provides continuity.\n\nPattern 4: AI Pair Programming (Aider/Claude Code)\n1. Start session with repo context\n2. Describe the task in natural language\n3. AI proposes changes as git diffs\n4. Review each diff before accepting\n5. AI commits with meaningful messages\n6. You handle edge cases and integration\n\nPattern 5: Autonomous Agent Workflow (OpenClaw/Claude Code)\n1. Define task in structured format (acceptance criteria, constraints)\n2. Agent plans → executes → verifies (reads files, runs tests)\n3. Agent creates PR/branch with changes\n4. You review the complete changeset\n5. Iterate on feedback\n\nPhase 6: Tool-Specific Power Moves\nCursor\nFeature\tPower Move\nTab completion\tLet it complete 3-5 tokens before accepting — catches wrong predictions early\nCmd+K (inline edit)\tSelect ONLY the exact lines to change — less context = more accurate\nChat\t@file to add context, @codebase for project-wide questions\nComposer\tMulti-file changes — describe the full feature, let it edit across files\n.cursorrules\tProject-specific AI instructions — commit to repo for team alignment\nNotepads\tReusable context (API docs, design docs) — attach to any chat\n\nCursor Pro Tips:\n\nUse @git to reference recent changes\nUse @docs to reference official library documentation\nCreate .cursor/rules/ directory for multiple rule files by domain\n\"Apply\" button to accept chat suggestions directly into code\nWindsurf (Cascade)\nFeature\tPower Move\nCascade flows\tMulti-step autonomous tasks — it can read, write, run terminal\nWrite mode\tDirect file editing with AI\nChat mode\tDiscussion without editing\n.windsurfrules\tProject context file\nTurbo mode\tFaster, less accurate — good for simple tasks\n\nWindsurf Pro Tips:\n\nCascade excels at multi-file refactors — give it the full scope\nUse \"undo flow\" to revert entire multi-step changes\nPin important files in context\nLet it read error output from terminal to self-fix\nCline\nFeature\tPower Move\nModel selection\tSwitch models per task (cheap for simple, expensive for complex)\nTool use\tReads files, runs commands, opens browser — full agent\nTransparency\tShows every action before executing — audit everything\nCustom instructions\tPer-project system prompts\nAuto-approve\tConfigure which actions need approval\n\nCline Pro Tips:\n\nSet spending limits to prevent runaway API costs\nUse cheaper models (Haiku/GPT-4o-mini) for simple tasks\nEnable \"diff mode\" to see exact changes before applying\nCreate task-specific instruction files\nAider\nFeature\tPower Move\n/add files\tExplicitly control which files the AI can see/edit\n/read files\tRead-only context (reference files)\n/architect\tTwo-model approach — architect plans, editor implements\nRepo map\tAuto-generates codebase summary for context\nGit integration\tEvery change is a commit — easy rollback\n\nAider Pro Tips:\n\nUse --architect flag for complex features (planner + implementer)\n/drop files you don't need to free context window\n--map-tokens to control repo map size\nRun aider --model claude-sonnet-4-20250514 for best code quality\nClaude Code\nFeature\tPower Move\nFull agent\tReads files, writes code, runs tests, git operations\nCLAUDE.md\tProject instructions file — auto-loaded\nSub-agents\tSpawn parallel workers for complex tasks\nMemory\tPersistent across sessions (project-level)\n\nClaude Code Pro Tips:\n\nWrite a comprehensive CLAUDE.md — it's your biggest leverage\nUse \"plan mode\" first for complex tasks, then \"implement\"\nLet it run tests and self-correct — don't interrupt the loop\nUse /compact when context gets long\nPhase 7: Code Quality Guardrails\nThe Trust-But-Verify Checklist\n\nAfter every AI-generated change:\n\n Read every line — don't blindly accept. AI hallucinates plausible-looking code\n Check imports — AI often imports non-existent modules or wrong versions\n Verify function signatures — parameter names, types, return types\n Test edge cases — AI optimizes for the happy path\n Check for security — hardcoded secrets, missing auth checks, SQL injection\n Run the tests — if tests pass, good. If no tests exist, write them first\n Check for drift — did it change files you didn't ask it to change?\n Verify dependencies — did it add packages? Are they real? Are they secure?\nCommon AI Code Failures\nFailure\tDetection\tFix\nHallucinated API\tCode uses functions that don't exist\tCheck library docs before accepting\nOutdated patterns\tUses deprecated APIs (React class components)\tSpecify versions in context\nMissing error handling\tHappy path only, no try/catch\tAsk specifically for error cases\nSecurity holes\tInline secrets, missing auth, XSS\tSecurity review as separate step\nOver-engineering\t5 files for a 20-line solution\tAsk for simplest possible solution\nWrong abstractions\tPremature generalization\tSpecify \"don't abstract, keep concrete\"\nTest theater\tTests that pass but test nothing\tReview test assertions specifically\nCopy-paste bugs\tDuplicated logic with subtle differences\tCheck for patterns, extract helpers\nThe 3-Read Review\nSkim read — Does the structure make sense? Right files, right approach?\nLogic read — Does each function do what it claims? Edge cases handled?\nIntegration read — Does it work with the rest of the codebase? Breaking changes?\nPhase 8: Cost Optimization\nToken Cost Awareness\nModel\tInput $/1M tokens\tOutput $/1M tokens\tBest For\nGPT-4o mini\t$0.15\t$0.60\tSimple completions, formatting\nClaude Haiku\t$0.25\t$1.25\tQuick edits, simple questions\nGPT-4o\t$2.50\t$10.00\tComplex code generation\nClaude Sonnet\t$3.00\t$15.00\tComplex code, long context\nClaude Opus\t$15.00\t$75.00\tArchitecture, hardest problems\no3\t$10.00\t$40.00\tComplex reasoning, algorithms\nCost Reduction Strategies\nTier your usage — Simple tasks → cheap model. Complex → expensive model\nReduce context — Every unnecessary file in context costs money\nStart new chats — Long conversations accumulate expensive history\nUse autocomplete for simple stuff — Copilot is flat-rate, much cheaper per completion\nCache project context — Use rules files instead of re-explaining every chat\nBatch related tasks — Handle related changes in one conversation\nMonthly Cost Benchmarks (Full-Time Developer)\nUsage Level\tEstimated Monthly Cost\nLight (Copilot + occasional chat)\t$20-40\nMedium (Cursor Pro + daily chat)\t$40-80\nHeavy (API-based agents, complex tasks)\t$80-200\nPower user (autonomous agents, all day)\t$200-500+\nPhase 9: Team Adoption\nRolling Out AI Coding Tools to a Team\n\nWeek 1-2: Foundation\n\nChoose primary tool (Cursor or Windsurf recommended for teams)\nCreate .cursorrules / .windsurfrules committed to repo\nRun a 1-hour workshop: basics, prompt techniques, verification\nSet team guidelines (review requirements, security rules)\n\nWeek 3-4: Practice\n\nDaily 15-min \"AI wins\" standup share\nPair sessions: experienced + new user\nCollect common prompts into team prompt library\nMonitor and address concerns (quality, dependency)\n\nMonth 2: Optimization\n\nMeasure: time-to-PR, bugs-per-feature, developer satisfaction\nIterate on .cursorrules based on team feedback\nCreate task-specific prompt templates in shared docs\nAddress skill gaps: who's using it well, who needs help?\n\nMonth 3: Systemization\n\nAI-assisted PR review as CI step\nAutomated test generation for new features\nCustom slash commands / snippets for team workflows\nQuarterly review: ROI, quality metrics, tooling updates\nTeam Guidelines Template\n# AI Coding Guidelines — [Team Name]\n\n## Approved Tools\n- [Tool 1] for [use case]\n- [Tool 2] for [use case]\n\n## Rules\n1. AI-generated code gets the SAME review rigor as human code\n2. Never paste proprietary/customer data into AI tools without approved data handling\n3. All AI-generated tests must be reviewed for assertion quality\n4. Security-sensitive code (auth, payments, PII) requires human-first approach\n5. Commit messages should NOT mention AI — own the code you commit\n\n## Quality Gates\n- [ ] Typecheck passes (`tsc --noEmit --strict`)\n- [ ] All tests pass\n- [ ] No new warnings\n- [ ] Manual review of all AI-generated code\n- [ ] Security-sensitive areas reviewed by security champion\n\nPhase 10: Advanced Patterns\nMulti-Agent Architecture for Development\nTask: Build feature X\n\nAgent 1 (Architect): Plans the approach, defines interfaces\nAgent 2 (Implementer): Writes the code\nAgent 3 (Tester): Writes and runs tests\nAgent 4 (Reviewer): Reviews for quality, security, patterns\n\nOrchestrator: Coordinates, resolves conflicts, maintains context\n\nSelf-Healing Development Loop\n1. Agent writes code\n2. Agent runs tests\n3. Tests fail → agent reads error, fixes code\n4. Repeat until tests pass\n5. Agent runs linter\n6. Lint fails → agent fixes\n7. All green → create PR\n\nThe Prompt Library Pattern\n\nMaintain a prompts/ directory in your project:\n\nprompts/\n  feature-implementation.md\n  bug-fix.md\n  refactoring.md\n  code-review.md\n  test-generation.md\n  migration.md\n  documentation.md\n\n\nEach file is a reusable prompt template. Reference them: \"Follow the template in prompts/feature-implementation.md\"\n\nModel Routing Strategy\ntask_routing:\n  autocomplete: copilot  # Always-on, flat rate\n  simple_edit: haiku     # Quick, cheap\n  feature_impl: sonnet   # Good balance\n  architecture: opus     # When it matters\n  debugging: sonnet      # Needs to reason about code\n  documentation: haiku   # Simple transformation\n  security_review: opus  # Can't afford mistakes\n  test_generation: sonnet # Needs understanding of code logic\n\nPhase 11: Anti-Patterns — What NOT to Do\nAnti-Pattern\tWhy It Fails\tDo This Instead\nPrompt and pray\tNo verification = bugs in production\tAlways review, always test\nPaste the whole codebase\tOverwhelms context, increases cost\tCurate relevant files only\nNever start new chats\tStale context → hallucinations\tNew task = new chat\nTrust without reading\tAI generates plausible but wrong code\tRead every line\nSkip tests because AI wrote it\tAI code has bugs too\tTest AI code MORE, not less\nUse one model for everything\tWaste money on simple tasks\tTier models by complexity\nNo project rules file\tAI guesses your conventions\tWrite .cursorrules / CLAUDE.md\nVague prompts\tGarbage in, garbage out\tUse SPEC framework\nOver-reliance\tSkill atrophy, can't debug AI output\tUnderstand what AI generates\nIgnoring security\tAI doesn't prioritize security\tExplicit security review step\nPhase 12: Scoring & Continuous Improvement\nAI-Assisted Development Quality Score (0-100)\nDimension\tWeight\tCriteria\nContext engineering\t20%\tRules files, curated context, fresh chats\nPrompt quality\t15%\tSPEC framework, task-appropriate templates\nVerification rigor\t20%\tReview checklist, test coverage, security review\nTool selection\t10%\tRight tool for task, model routing\nCost efficiency\t10%\tTiered usage, context management, batch tasks\nOutput quality\t15%\tCode correctness, maintainability, no drift\nWorkflow integration\t10%\tSystematic process, team alignment\nWeekly Self-Review Questions\nWhat was my best AI-assisted output this week? What made it good?\nWhere did AI waste my time? What went wrong with context/prompts?\nAm I reviewing thoroughly enough, or rubber-stamping?\nWhat prompt patterns worked well? Add to prompt library.\nAm I over-relying on AI for things I should understand deeply?\nMonthly Metrics\nAcceleration factor: Tasks completed per day vs pre-AI baseline\nBug rate: Bugs in AI-assisted code vs manual code\nCost per feature: API spend / features shipped\nContext efficiency: Average conversation length before drift\nCoverage: % of codebase with AI-assisted tests\nQuick Reference: Natural Language Commands\n\"Set up AI coding for [project]\" — Generate rules file + tool recommendations\n\"Write a prompt for [task type]\" — Generate SPEC-formatted prompt template\n\"Review this AI output\" — Run the Trust-But-Verify checklist\n\"Compare [tool A] vs [tool B] for [use case]\" — Tool selection analysis\n\"Optimize my AI coding costs\" — Analyze usage and suggest model routing\n\"Create a team AI coding guide\" — Generate team guidelines document\n\"Debug why AI keeps [hallucinating X]\" — Context diagnosis\n\"Set up test-driven AI workflow for [feature]\" — TDD-AI pattern guide\n\"Create prompt library for [project type]\" — Generate prompt templates\n\"Score my AI coding maturity\" — Run the quality assessment\n\"Onboard [person] to AI coding\" — Generate training plan\n\"Audit AI coding security practices\" — Security review checklist"
  },
  "trust": {
    "sourceLabel": "tencent",
    "provenanceUrl": "https://clawhub.ai/1kalin/afrexai-ai-coding-toolkit",
    "publisherUrl": "https://clawhub.ai/1kalin/afrexai-ai-coding-toolkit",
    "owner": "1kalin",
    "version": "1.0.0",
    "license": null,
    "verificationStatus": "Indexed source record"
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/afrexai-ai-coding-toolkit",
    "downloadUrl": "https://openagent3.xyz/downloads/afrexai-ai-coding-toolkit",
    "agentUrl": "https://openagent3.xyz/skills/afrexai-ai-coding-toolkit/agent",
    "manifestUrl": "https://openagent3.xyz/skills/afrexai-ai-coding-toolkit/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/afrexai-ai-coding-toolkit/agent.md"
  }
}