$npx -y skills add Aedelon/claude-code-blueprint --skill core-protocolsDebug errors systematically by searching first, then analyzing, then proposing verified solutions. MUST BE USED when user reports: "error", "bug", "doesn't work", "fails", "crash", stack traces, exception messages, or any troubleshooting scenario. Triggers: "TypeError", "ImportEr
| 1 | # Core Protocols |
| 2 | |
| 3 | ## 1. Debugging Protocol |
| 4 | |
| 5 | ### Trigger Conditions |
| 6 | |
| 7 | Activate when: |
| 8 | - User reports an error or bug |
| 9 | - Code doesn't work as expected |
| 10 | - Build/test failures occur |
| 11 | - Stack traces or error messages appear |
| 12 | |
| 13 | ### MANDATORY Sequence |
| 14 | |
| 15 | **NEVER propose solutions before completing Steps 1-2.** |
| 16 | |
| 17 | #### Step 1: SEARCH FIRST |
| 18 | |
| 19 | ``` |
| 20 | 1. Extract EXACT error message |
| 21 | 2. WebSearch: "[error message] [language/framework] [version]" |
| 22 | 3. Context7 if library-related: |
| 23 | - mcp__context7__resolve-library-id(libraryName) |
| 24 | - mcp__context7__get-library-docs(id, topic: "error") |
| 25 | 4. Look for GitHub issues / StackOverflow |
| 26 | ``` |
| 27 | |
| 28 | #### Step 2: ANALYZE |
| 29 | |
| 30 | ``` |
| 31 | 1. Identify ROOT CAUSE (not symptoms) |
| 32 | 2. Check version compatibility |
| 33 | 3. Isolate the problem |
| 34 | 4. Check changelogs for breaking changes |
| 35 | ``` |
| 36 | |
| 37 | #### Step 3: VERIFIED SOLUTION |
| 38 | |
| 39 | ``` |
| 40 | 1. Cite SOURCE of solution |
| 41 | 2. State CONFIDENCE level |
| 42 | 3. Propose ALTERNATIVES if uncertain |
| 43 | 4. Include VALIDATION command |
| 44 | ``` |
| 45 | |
| 46 | ### Error Type Quick Reference |
| 47 | |
| 48 | | Error Type | First Action | |
| 49 | |------------|--------------| |
| 50 | | Library/API error | Context7 lookup | |
| 51 | | Runtime error | WebSearch exact message | |
| 52 | | Build error | Check versions, config | |
| 53 | | Type error | Read type definitions | |
| 54 | | Import error | Check package installation | |
| 55 | |
| 56 | --- |
| 57 | |
| 58 | ## 2. Confidence Assessment |
| 59 | |
| 60 | ### Levels |
| 61 | |
| 62 | | Level | Criteria | Action | |
| 63 | |-------|----------|--------| |
| 64 | | **HIGH** | Verified via tool, 2+ sources | State source | |
| 65 | | **MEDIUM** | Single source, partial verification | Add caveat | |
| 66 | | **LOW** | Memory only, no verification | Warn explicitly | |
| 67 | | **UNKNOWN** | Cannot assess | Say "I don't know" | |
| 68 | |
| 69 | ### Verification Methods |
| 70 | |
| 71 | **Libraries/APIs**: Context7 → cross-reference docs → HIGH if verified |
| 72 | **Current Events**: WebSearch 2024-2025 → multiple sources → HIGH if recent |
| 73 | **Best Practices**: Official docs → community consensus → MEDIUM-HIGH |
| 74 | |
| 75 | ### Mandatory Warnings |
| 76 | |
| 77 | | Topic | Warning | |
| 78 | |-------|---------| |
| 79 | | Pricing | "Verify at official URL" | |
| 80 | | Rate limits | "Check your dashboard" | |
| 81 | | New features | "As of [DATE], may have changed" | |
| 82 | | Model IDs | "Verify current availability" | |
| 83 | |
| 84 | --- |
| 85 | |
| 86 | ## 3. Output Templates |
| 87 | |
| 88 | ### Code Response |
| 89 | |
| 90 | ```markdown |
| 91 | **Intent**: [1-2 sentences] |
| 92 | |
| 93 | \`\`\`[language] |
| 94 | [code with comments] |
| 95 | \`\`\` |
| 96 | |
| 97 | **Validation**: `[command]` |
| 98 | **Dependencies**: [if new] |
| 99 | **Confidence**: [LEVEL] - [source] |
| 100 | ``` |
| 101 | |
| 102 | ### Research Response |
| 103 | |
| 104 | ```markdown |
| 105 | ## Summary |
| 106 | [3-5 sentences max] |
| 107 | |
| 108 | ## Sources |
| 109 | 1. [Author (Year). Title. DOI/URL] |
| 110 | |
| 111 | **Confidence**: [LEVEL] - [justification] |
| 112 | **Limitations**: [what couldn't be verified] |
| 113 | ``` |
| 114 | |
| 115 | ### Diagnostic Response |
| 116 | |
| 117 | ```markdown |
| 118 | ## Problem Identified |
| 119 | [1-2 sentences] |
| 120 | |
| 121 | ## Root Cause |
| 122 | [Technical explanation] |
| 123 | **Source**: [reference] |
| 124 | |
| 125 | ## Solution |
| 126 | \`\`\`[language] |
| 127 | [fix] |
| 128 | \`\`\` |
| 129 | |
| 130 | ## Validation |
| 131 | \`\`\`bash |
| 132 | [command] |
| 133 | \`\`\` |
| 134 | |
| 135 | **Confidence**: [LEVEL] |
| 136 | **Prevention**: [how to avoid] |
| 137 | ``` |
| 138 | |
| 139 | ### Creative Prompt Response |
| 140 | |
| 141 | ```markdown |
| 142 | **Prompt [Tool]**: |
| 143 | [ready-to-use prompt] |
| 144 | |
| 145 | **Parameters**: [key settings] |
| 146 | **Variations**: [alternatives] |
| 147 | ``` |
| 148 | |
| 149 | --- |
| 150 | |
| 151 | ## Anti-Patterns (FORBIDDEN) |
| 152 | |
| 153 | - Proposing solutions without searching first |
| 154 | - Guessing based on similar-sounding errors |
| 155 | - Omitting confidence levels |
| 156 | - Skipping validation commands |
| 157 | - Ignoring version information |