[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6D347F7K66K9SHRK9WZFKA7":3,"skill-same:01M6D347F7K66K9SHRK9WZFKA7":70},{"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":24,"linkOnly":25,"markdownContent":26,"markdownBody":27,"frontMatter":28,"refName":29,"commitSha":30,"skillDirectory":31,"forksCount":32,"repositoryHtmlUrl":33,"repositoryHomepageUrl":6,"ownerHtmlUrl":34,"ownerBlogUrl":35,"ownerDescription":36,"updatedAt":37,"install":38,"resources":51},"01M6D347F7K66K9SHRK9WZFKA7","amp-improve",null,"Workshop an improvement idea into implementation plan","hemory\u002Famp-legacy-do-not-use","hemory","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F21179881?v=4","01M6D31HFJ391DN8N0JP8AQGFE","Hemory Phifer","01M6D31HKXVTEN27AT6R4PP2ZJ",3,"MIT","https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Famp-improve\u002FSKILL.md",[],[19,20,21,22,23],"improve","amp improve","plan","implementation","idea","2026-11-10 10:12:30",false,"---\nname: amp-improve\ndescription: Workshop an improvement idea into implementation plan\n---\n\n## What This Command Does\n\n**In plain English:** Your design partner for making Amp better. Three ways to use it:\n\n1. **Workshop an idea** — Bring a fuzzy improvement idea, we'll shape it into a concrete plan\n2. **Review what's new** — See recent Claude Code changes and how they could improve your system\n3. **Full capability audit** — Find features you're not using that could help with your jobs-to-be-done\n\n**When to use it:**\n- \"I have an idea but don't know how to implement it\"\n- \"What new Claude Code features should I be using?\"\n- \"Am I getting the most out of this system?\"\n\n**How to run it:**\n```\n\u002Famp-improve                           # Shows menu of options\n\u002Famp-improve \"your idea here\"          # Jumps straight to workshopping\n```\n\n---\n\n## Entry Point\n\nWhen invoked, offer these options:\n\n```\nHow would you like to explore improvements?\n\n1. **I have an idea** — Workshop a specific improvement\n2. **What's new?** — Review Claude Code changes since [last check: DATE]\n3. **Full audit** — Review all capabilities vs. your current usage\n\nOr just describe your idea and I'll start workshopping.\n```\n\nIf user provides $IDEA argument, skip to Mode 1 directly.\n\n---\n\n## Mode 1: Workshop an Idea\n\n**Arguments:** $IDEA — High-level description of what you want to improve\n\n### Phase 1: Understand\n\nParse the idea and frame it back:\n\n1. Identify which Amp areas are involved:\n   - Workflows (daily, weekly routines)\n   - Automation (scripts, hooks, scheduled tasks)\n   - Relationships (people tracking, meetings)\n   - Task management (capture, flow, priorities)\n   - Projects (planning, tracking, health)\n   - Information retrieval (search, lookups)\n\n2. State your understanding in 2-3 sentences\n3. Note any ambiguity that needs clarification\n\n### Phase 2: Research + Capability Scan\n\nBuild context from multiple sources:\n\n**Internal (always check):**\n```\n06-Resources\u002FLearnings\u002F                 # What you've learned so far\nSystem\u002FSkills\u002F                       # Existing reusable behaviors\n.claude\u002Fcommands\u002F                    # Existing commands that might overlap\nSystem\u002Fpillars.yaml                  # Strategic priorities\n```\n\n**External (when needed):**\n- Search web for latest Claude Code capabilities\n- Check official docs at docs.anthropic.com\n\nCheck:\n- Does similar functionality already exist?\n- What existing patterns could we leverage?\n\n### Phase 3: Capability Match\n\nMap requirements to optimal Claude Code features:\n\n| Requirement Pattern | Suggested Feature | Why |\n|---------------------|-------------------|-----|\n| \"Every time I edit X, do Y\" | PostToolUse hook | Automatic trigger after tool completion |\n| \"Before doing X, check Y\" | PreToolUse hook | Validation\u002Fconfirmation gate |\n| \"At start of session, load Z\" | SessionStart hook | Context injection |\n| \"Fast search while I work\" | Explore sub-agent | Haiku model, runs in parallel |\n| \"Different perspective on X\" | Custom sub-agent | Isolated context, specific focus |\n| \"Multi-step structured workflow\" | Skill | Reusable behavior, always available |\n| \"Explicit invocation with steps\" | Command | User-triggered workflow |\n| \"Connect to external service\" | MCP integration | Tool access to external systems |\n\nPresent the capability analysis to the user.\n\n### Phase 4: Expand\n\nSuggest 2-3 adjacent improvements:\n\n- Based on capability analysis, what else could we solve?\n- Flag compound opportunities (build A, get B nearly free)\n- Connect to existing gaps in the system\n- Consider synergies with strategic pillars\n\nFrame as: \"While we're here, you might also consider...\"\n\n### Phase 5: Refine\n\nDrive toward concrete requirements through focused questions:\n\n- Ask max 2-3 questions at a time\n- Validate capability assumptions\n- Confirm implementation approach preference\n- Identify acceptance criteria\n\nLoop between Research → Expand → Refine until requirements are solid.\n\n### Phase 6: Plan + Update\n\nWhen requirements are firm:\n\n1. Generate the plan document (see template below)\n2. Save to `plans\u002Famp-improvement-[slug].md`\n3. Add any new patterns to `06-Resources\u002FLearnings\u002F`\n4. Add viable future ideas to backlog\n\n---\n\n## Mode 2: What's New?\n\nReview recent Claude Code changes and suggest improvements.\n\n### Process\n\n1. **Read state file:** `System\u002Fclaude-code-state.json`\n   - If missing, this is first run — do a full scan instead\n\n2. **Fetch current changelog:** Use WebSearch to find latest Claude Code changelog\u002Freleases\n\n3. **Compare:** What's new since `last_check` date?\n\n4. **For each new feature, evaluate:**\n   - Is this relevant to Amp use cases?\n   - What existing workflow could it improve?\n   - Does it unlock something previously impossible?\n\n5. **Present findings:**\n\n```\n? Claude Code Updates (since [last check date])\n\n=== NEW CAPABILITIES ===\n\n1. [Feature Name]\n   - What it does: [description]\n   - How it could help: [specific Amp improvement]\n   - Effort to implement: Low\u002FMedium\u002FHigh\n\n2. [Feature Name]\n   ...\n\n=== SUGGESTED IMPROVEMENTS ===\n\nBased on these new capabilities, here's what you could build:\n\n1. [Improvement name]\n   - What: [description]\n   - Why: [connects to which pillar\u002Fjob-to-be-done]\n   - Uses: [which new capability]\n\nWant to workshop any of these? (Enter number or describe your own idea)\n```\n\n6. **Update state file** with current date and capabilities seen\n\n### State File Format\n\nLocation: `System\u002Fclaude-code-state.json`\n\n```json\n{\n  \"last_check\": \"2026-01-22\",\n  \"last_version_seen\": \"1.0.30\",\n  \"features_seen\": [\n    \"hooks\",\n    \"sub-agents\",\n    \"skills\",\n    \"commands\",\n    \"mcp\"\n  ],\n  \"features_noted\": [\n    {\n      \"version\": \"1.0.30\",\n      \"feature\": \"Async hooks\",\n      \"date_seen\": \"2026-01-22\"\n    }\n  ]\n}\n```\n\n---\n\n## Mode 3: Full Capability Audit\n\nReview all known Claude Code capabilities against your current system.\n\n### Process\n\n1. **Inventory current usage:**\n   - Scan `.claude\u002F` folder for hooks, commands, settings\n   - Check `System\u002FSkills\u002F` for skills\n   - Check `core\u002Fmcp\u002F` for MCP integrations\n   - Read `06-Resources\u002FLearnings\u002F` for known patterns\n\n2. **Load capability reference:**\n   - Use the capability match table (from Mode 1, Phase 3)\n   - Search web for comprehensive Claude Code feature list if needed\n\n3. **Gap analysis:**\n\n```\n=== CAPABILITY AUDIT ===\n\n✅ Using Well:\n- Commands (12 defined)\n- Skills (person-lookup)\n\n⚠️ Underutilized:\n- Hooks — No hooks configured. Could automate:\n  - Session start: Load context automatically\n  - Post-edit: Update related files\n\n❌ Not Using:\n- Sub-agents — Could parallelize research tasks\n- MCP beyond tasks — Calendar, email integrations possible\n\n=== RECOMMENDATIONS ===\n\nRanked by impact vs. effort:\n\n1. [High impact, low effort]\n   - Add SessionStart hook to load daily context\n   - Why: Saves 30 seconds every session\n   - Connects to: [pillar]\n\n2. [High impact, medium effort]\n   ...\n```\n\n4. **Offer to workshop** any recommendation the user wants to pursue\n\n---\n\n## Plan Output Template\n\nCreate at: `plans\u002Famp-improvement-[slug].md`\n\n```markdown\n# Amp Improvement: [Title]\n\n**Created:** [Date]\n**Status:** Planning → In Progress → Complete\n**Pillar:** [pillar name if configured]\n\n## Overview\n\n[2-3 sentence summary of what we're improving and why]\n\n## Requirements\n\n- [ ] Requirement 1\n- [ ] Requirement 2\n- [ ] Requirement 3\n\n## Capability Analysis\n\n| Requirement | Implementation | Feature Type | Rationale |\n|-------------|----------------|--------------|-----------|\n| [req] | [how] | Hook\u002FSub-agent\u002FSkill\u002FCommand | [why this approach] |\n\n## Recommended Approach\n\n[Which Claude Code features to use and why. Include trade-offs considered.]\n\n## Implementation Steps\n\n### Phase 1: [Foundation]\n1. [Step]\n2. [Step]\n\n### Phase 2: [Core]\n1. [Step]\n2. [Step]\n\n## Files to Create\u002FModify\n\n| File | Action | Purpose |\n|------|--------|---------|\n| `path\u002Fto\u002Ffile` | Create\u002FModify | [what] |\n\n## Compound Opportunities\n\n[What else we could solve while building this]\n\n- [ ] [Adjacent improvement 1]\n- [ ] [Adjacent improvement 2]\n\n## Acceptance Criteria\n\n- [ ] [Testable outcome 1]\n- [ ] [Testable outcome 2]\n\n## Questions Resolved\n\n- Q: [Question that came up]\n- A: [How we decided]\n```\n\n---\n\n## Example Interactions\n\n### Mode 1: Workshop\n```\n\u002Famp-improve \"Better meeting action item capture\"\n\n> **Understanding:** You want to improve how action items flow from meetings \n> into your task system. This touches workflows (meeting processing) and \n> task management (capture and routing).\n>\n> **Current state** (from Learnings):\n> - Meeting notes go to `00-Inbox\u002FMeetings\u002F`\n> - Action items manually extracted during review\n>\n> **Capability scan:**\n> - PostToolUse hook could auto-trigger processing after meeting note creation\n> - Sub-agent could parallel process: action items, follow-ups, person updates\n>\n> **Questions:**\n> 1. Where are items getting lost—during capture, or in follow-through?\n> 2. Would auto-drafted follow-up emails be valuable?\n```\n\n### Mode 2: What's New\n```\n\u002Famp-improve\n> [selects \"What's new?\"]\n\n> ? Claude Code Updates (since 2026-01-15)\n>\n> === NEW CAPABILITIES ===\n> 1. Async Hooks — Hooks can now run without blocking\n>    - How it could help: Faster session startup\n>    - Effort: Low\n>\n> === SUGGESTED IMPROVEMENTS ===\n> 1. Non-blocking context loading\n>    - What: Load learnings\u002Fpreferences in background\n>    - Why: Snappier session starts\n>\n> Want to workshop any of these?\n```\n\n### Mode 3: Full Audit\n```\n\u002Famp-improve\n> [selects \"Full audit\"]\n\n> === CAPABILITY AUDIT ===\n>\n> ✅ Using Well: Commands (12), Skills (1)\n> ⚠️ Underutilized: Hooks (0 configured)\n> ❌ Not Using: Sub-agents, Advanced MCP\n>\n> === TOP RECOMMENDATION ===\n> Add SessionStart hook — saves context loading time every session\n>\n> Want me to workshop this improvement?\n```\n\n---\n\n## Behaviors\n\n### Always Do\n- Check internal patterns before suggesting new implementations\n- Tie suggestions to user's pillars\u002Fjobs-to-be-done when configured\n- Update state file after Mode 2 runs\n- Offer concrete capability recommendations, not just ideas\n\n### Never Do\n- Skip the research phase\n- Suggest capabilities without explaining why they fit\n- Output a plan before requirements are confirmed\n- Make changes without explicit approval\n\n---\n\n## Philosophy\n\nThis isn't just requirements gathering—it's capability-aware design. The goal is to help you leverage Claude Code's full potential while building exactly what you need, nothing more.\n\n---\n\n## Track Usage (Silent)\n\nUpdate `System\u002Fusage_log.md` to mark improvement workshop as used.\n\n**Analytics (Silent):**\n\nCall `track_event` with event_name `improvement_workshopped` and properties:\n- `idea_id`\n\nThis only fires if the user has opted into analytics. No action needed if it returns \"analytics_disabled\".\n","## What This Command Does\n\n**In plain English:** Your design partner for making Amp better. Three ways to use it:\n\n1. **Workshop an idea** — Bring a fuzzy improvement idea, we'll shape it into a concrete plan\n2. **Review what's new** — See recent Claude Code changes and how they could improve your system\n3. **Full capability audit** — Find features you're not using that could help with your jobs-to-be-done\n\n**When to use it:**\n- \"I have an idea but don't know how to implement it\"\n- \"What new Claude Code features should I be using?\"\n- \"Am I getting the most out of this system?\"\n\n**How to run it:**\n```\n\u002Famp-improve                           # Shows menu of options\n\u002Famp-improve \"your idea here\"          # Jumps straight to workshopping\n```\n\n---\n\n## Entry Point\n\nWhen invoked, offer these options:\n\n```\nHow would you like to explore improvements?\n\n1. **I have an idea** — Workshop a specific improvement\n2. **What's new?** — Review Claude Code changes since [last check: DATE]\n3. **Full audit** — Review all capabilities vs. your current usage\n\nOr just describe your idea and I'll start workshopping.\n```\n\nIf user provides $IDEA argument, skip to Mode 1 directly.\n\n---\n\n## Mode 1: Workshop an Idea\n\n**Arguments:** $IDEA — High-level description of what you want to improve\n\n### Phase 1: Understand\n\nParse the idea and frame it back:\n\n1. Identify which Amp areas are involved:\n   - Workflows (daily, weekly routines)\n   - Automation (scripts, hooks, scheduled tasks)\n   - Relationships (people tracking, meetings)\n   - Task management (capture, flow, priorities)\n   - Projects (planning, tracking, health)\n   - Information retrieval (search, lookups)\n\n2. State your understanding in 2-3 sentences\n3. Note any ambiguity that needs clarification\n\n### Phase 2: Research + Capability Scan\n\nBuild context from multiple sources:\n\n**Internal (always check):**\n```\n06-Resources\u002FLearnings\u002F                 # What you've learned so far\nSystem\u002FSkills\u002F                       # Existing reusable behaviors\n.claude\u002Fcommands\u002F                    # Existing commands that might overlap\nSystem\u002Fpillars.yaml                  # Strategic priorities\n```\n\n**External (when needed):**\n- Search web for latest Claude Code capabilities\n- Check official docs at docs.anthropic.com\n\nCheck:\n- Does similar functionality already exist?\n- What existing patterns could we leverage?\n\n### Phase 3: Capability Match\n\nMap requirements to optimal Claude Code features:\n\n| Requirement Pattern | Suggested Feature | Why |\n|---------------------|-------------------|-----|\n| \"Every time I edit X, do Y\" | PostToolUse hook | Automatic trigger after tool completion |\n| \"Before doing X, check Y\" | PreToolUse hook | Validation\u002Fconfirmation gate |\n| \"At start of session, load Z\" | SessionStart hook | Context injection |\n| \"Fast search while I work\" | Explore sub-agent | Haiku model, runs in parallel |\n| \"Different perspective on X\" | Custom sub-agent | Isolated context, specific focus |\n| \"Multi-step structured workflow\" | Skill | Reusable behavior, always available |\n| \"Explicit invocation with steps\" | Command | User-triggered workflow |\n| \"Connect to external service\" | MCP integration | Tool access to external systems |\n\nPresent the capability analysis to the user.\n\n### Phase 4: Expand\n\nSuggest 2-3 adjacent improvements:\n\n- Based on capability analysis, what else could we solve?\n- Flag compound opportunities (build A, get B nearly free)\n- Connect to existing gaps in the system\n- Consider synergies with strategic pillars\n\nFrame as: \"While we're here, you might also consider...\"\n\n### Phase 5: Refine\n\nDrive toward concrete requirements through focused questions:\n\n- Ask max 2-3 questions at a time\n- Validate capability assumptions\n- Confirm implementation approach preference\n- Identify acceptance criteria\n\nLoop between Research → Expand → Refine until requirements are solid.\n\n### Phase 6: Plan + Update\n\nWhen requirements are firm:\n\n1. Generate the plan document (see template below)\n2. Save to `plans\u002Famp-improvement-[slug].md`\n3. Add any new patterns to `06-Resources\u002FLearnings\u002F`\n4. Add viable future ideas to backlog\n\n---\n\n## Mode 2: What's New?\n\nReview recent Claude Code changes and suggest improvements.\n\n### Process\n\n1. **Read state file:** `System\u002Fclaude-code-state.json`\n   - If missing, this is first run — do a full scan instead\n\n2. **Fetch current changelog:** Use WebSearch to find latest Claude Code changelog\u002Freleases\n\n3. **Compare:** What's new since `last_check` date?\n\n4. **For each new feature, evaluate:**\n   - Is this relevant to Amp use cases?\n   - What existing workflow could it improve?\n   - Does it unlock something previously impossible?\n\n5. **Present findings:**\n\n```\n? Claude Code Updates (since [last check date])\n\n=== NEW CAPABILITIES ===\n\n1. [Feature Name]\n   - What it does: [description]\n   - How it could help: [specific Amp improvement]\n   - Effort to implement: Low\u002FMedium\u002FHigh\n\n2. [Feature Name]\n   ...\n\n=== SUGGESTED IMPROVEMENTS ===\n\nBased on these new capabilities, here's what you could build:\n\n1. [Improvement name]\n   - What: [description]\n   - Why: [connects to which pillar\u002Fjob-to-be-done]\n   - Uses: [which new capability]\n\nWant to workshop any of these? (Enter number or describe your own idea)\n```\n\n6. **Update state file** with current date and capabilities seen\n\n### State File Format\n\nLocation: `System\u002Fclaude-code-state.json`\n\n```json\n{\n  \"last_check\": \"2026-01-22\",\n  \"last_version_seen\": \"1.0.30\",\n  \"features_seen\": [\n    \"hooks\",\n    \"sub-agents\",\n    \"skills\",\n    \"commands\",\n    \"mcp\"\n  ],\n  \"features_noted\": [\n    {\n      \"version\": \"1.0.30\",\n      \"feature\": \"Async hooks\",\n      \"date_seen\": \"2026-01-22\"\n    }\n  ]\n}\n```\n\n---\n\n## Mode 3: Full Capability Audit\n\nReview all known Claude Code capabilities against your current system.\n\n### Process\n\n1. **Inventory current usage:**\n   - Scan `.claude\u002F` folder for hooks, commands, settings\n   - Check `System\u002FSkills\u002F` for skills\n   - Check `core\u002Fmcp\u002F` for MCP integrations\n   - Read `06-Resources\u002FLearnings\u002F` for known patterns\n\n2. **Load capability reference:**\n   - Use the capability match table (from Mode 1, Phase 3)\n   - Search web for comprehensive Claude Code feature list if needed\n\n3. **Gap analysis:**\n\n```\n=== CAPABILITY AUDIT ===\n\n✅ Using Well:\n- Commands (12 defined)\n- Skills (person-lookup)\n\n⚠️ Underutilized:\n- Hooks — No hooks configured. Could automate:\n  - Session start: Load context automatically\n  - Post-edit: Update related files\n\n❌ Not Using:\n- Sub-agents — Could parallelize research tasks\n- MCP beyond tasks — Calendar, email integrations possible\n\n=== RECOMMENDATIONS ===\n\nRanked by impact vs. effort:\n\n1. [High impact, low effort]\n   - Add SessionStart hook to load daily context\n   - Why: Saves 30 seconds every session\n   - Connects to: [pillar]\n\n2. [High impact, medium effort]\n   ...\n```\n\n4. **Offer to workshop** any recommendation the user wants to pursue\n\n---\n\n## Plan Output Template\n\nCreate at: `plans\u002Famp-improvement-[slug].md`\n\n```markdown\n# Amp Improvement: [Title]\n\n**Created:** [Date]\n**Status:** Planning → In Progress → Complete\n**Pillar:** [pillar name if configured]\n\n## Overview\n\n[2-3 sentence summary of what we're improving and why]\n\n## Requirements\n\n- [ ] Requirement 1\n- [ ] Requirement 2\n- [ ] Requirement 3\n\n## Capability Analysis\n\n| Requirement | Implementation | Feature Type | Rationale |\n|-------------|----------------|--------------|-----------|\n| [req] | [how] | Hook\u002FSub-agent\u002FSkill\u002FCommand | [why this approach] |\n\n## Recommended Approach\n\n[Which Claude Code features to use and why. Include trade-offs considered.]\n\n## Implementation Steps\n\n### Phase 1: [Foundation]\n1. [Step]\n2. [Step]\n\n### Phase 2: [Core]\n1. [Step]\n2. [Step]\n\n## Files to Create\u002FModify\n\n| File | Action | Purpose |\n|------|--------|---------|\n| `path\u002Fto\u002Ffile` | Create\u002FModify | [what] |\n\n## Compound Opportunities\n\n[What else we could solve while building this]\n\n- [ ] [Adjacent improvement 1]\n- [ ] [Adjacent improvement 2]\n\n## Acceptance Criteria\n\n- [ ] [Testable outcome 1]\n- [ ] [Testable outcome 2]\n\n## Questions Resolved\n\n- Q: [Question that came up]\n- A: [How we decided]\n```\n\n---\n\n## Example Interactions\n\n### Mode 1: Workshop\n```\n\u002Famp-improve \"Better meeting action item capture\"\n\n> **Understanding:** You want to improve how action items flow from meetings \n> into your task system. This touches workflows (meeting processing) and \n> task management (capture and routing).\n>\n> **Current state** (from Learnings):\n> - Meeting notes go to `00-Inbox\u002FMeetings\u002F`\n> - Action items manually extracted during review\n>\n> **Capability scan:**\n> - PostToolUse hook could auto-trigger processing after meeting note creation\n> - Sub-agent could parallel process: action items, follow-ups, person updates\n>\n> **Questions:**\n> 1. Where are items getting lost—during capture, or in follow-through?\n> 2. Would auto-drafted follow-up emails be valuable?\n```\n\n### Mode 2: What's New\n```\n\u002Famp-improve\n> [selects \"What's new?\"]\n\n> ? Claude Code Updates (since 2026-01-15)\n>\n> === NEW CAPABILITIES ===\n> 1. Async Hooks — Hooks can now run without blocking\n>    - How it could help: Faster session startup\n>    - Effort: Low\n>\n> === SUGGESTED IMPROVEMENTS ===\n> 1. Non-blocking context loading\n>    - What: Load learnings\u002Fpreferences in background\n>    - Why: Snappier session starts\n>\n> Want to workshop any of these?\n```\n\n### Mode 3: Full Audit\n```\n\u002Famp-improve\n> [selects \"Full audit\"]\n\n> === CAPABILITY AUDIT ===\n>\n> ✅ Using Well: Commands (12), Skills (1)\n> ⚠️ Underutilized: Hooks (0 configured)\n> ❌ Not Using: Sub-agents, Advanced MCP\n>\n> === TOP RECOMMENDATION ===\n> Add SessionStart hook — saves context loading time every session\n>\n> Want me to workshop this improvement?\n```\n\n---\n\n## Behaviors\n\n### Always Do\n- Check internal patterns before suggesting new implementations\n- Tie suggestions to user's pillars\u002Fjobs-to-be-done when configured\n- Update state file after Mode 2 runs\n- Offer concrete capability recommendations, not just ideas\n\n### Never Do\n- Skip the research phase\n- Suggest capabilities without explaining why they fit\n- Output a plan before requirements are confirmed\n- Make changes without explicit approval\n\n---\n\n## Philosophy\n\nThis isn't just requirements gathering—it's capability-aware design. The goal is to help you leverage Claude Code's full potential while building exactly what you need, nothing more.\n\n---\n\n## Track Usage (Silent)\n\nUpdate `System\u002Fusage_log.md` to mark improvement workshop as used.\n\n**Analytics (Silent):**\n\nCall `track_event` with event_name `improvement_workshopped` and properties:\n- `idea_id`\n\nThis only fires if the user has opted into analytics. No action needed if it returns \"analytics_disabled\".\n",{"name":5,"description":7},"main","5417cad9215b767627c7c947dcb3a0bbe3ebe52d",".claude\u002Fskills\u002Famp-improve",0,"https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use","https:\u002F\u002Fgithub.com\u002Fhemory","https:\u002F\u002Fwww.heminthehat.com\u002F","The People Developer | L&D Program Manager | Software Engineer","2026-06-04 23:46:30",{"source":8,"skillName":5,"commands":39,"prompt":49,"zipAvailable":50},[40,43,46],{"tool":41,"command":42},"npx","npx skills add hemory\u002Famp-legacy-do-not-use --skill amp-improve",{"tool":44,"command":45},"bunx","bunx skills add hemory\u002Famp-legacy-do-not-use --skill amp-improve",{"tool":47,"command":48},"pnpm","pnpm dlx skills add hemory\u002Famp-legacy-do-not-use --skill amp-improve","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Famp-improve\u002FSKILL.md",true,[52,58,62,66],{"path":53,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":56,"sourceHtmlUrl":57},"references\u002Fadapter-system.md","REFERENCE","参考文档",9468,"https:\u002F\u002Fgithub.com\u002Ffadrienne\u002FMyGitHub\u002Fblob\u002Fmain\u002Fopenscience\u002Fbackend\u002Fcli\u002Fskills\u002Fml-training\u002Fnemo-evaluator\u002Freferences\u002Fadapter-system.md",{"path":59,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":60,"sourceHtmlUrl":61},"references\u002Fconfiguration.md",9567,"https:\u002F\u002Fgithub.com\u002Ffadrienne\u002FMyGitHub\u002Fblob\u002Fmain\u002Fopenscience\u002Fbackend\u002Fcli\u002Fskills\u002Fml-training\u002Fnemo-evaluator\u002Freferences\u002Fconfiguration.md",{"path":63,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":64,"sourceHtmlUrl":65},"references\u002Fcustom-benchmarks.md",7101,"https:\u002F\u002Fgithub.com\u002Ffadrienne\u002FMyGitHub\u002Fblob\u002Fmain\u002Fopenscience\u002Fbackend\u002Fcli\u002Fskills\u002Fml-training\u002Fnemo-evaluator\u002Freferences\u002Fcustom-benchmarks.md",{"path":67,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":68,"sourceHtmlUrl":69},"references\u002Fexecution-backends.md",9351,"https:\u002F\u002Fgithub.com\u002Ffadrienne\u002FMyGitHub\u002Fblob\u002Fmain\u002Fopenscience\u002Fbackend\u002Fcli\u002Fskills\u002Fml-training\u002Fnemo-evaluator\u002Freferences\u002Fexecution-backends.md",[71,108,117,131,145,157],{"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":87,"publishedAt":24},"01M6D3DVEAHJTTK4W21B15ZZKK","zoom-setup","Connect Zoom to Amp for meeting recordings, scheduling, and transcript context","https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fzoom-setup\u002FSKILL.md",[77,82],{"categoryNo":78,"categoryName":79,"taxonomyType":80,"taxonomyTypeDesc":81},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",{"categoryNo":83,"categoryName":84,"taxonomyType":85,"taxonomyTypeDesc":86},"01JBSEEDCAT000000000000034","行政与办公人员","PROFESSION","适用职业",[88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107],"zoom setup","setup","schedule","MCP","review","connect","transcript","search","stats","oauth2","confirmation","context","meetings","process","recordings","surfaces","scheduling","granola","time","distribution",{"skillNo":109,"name":110,"nameZh":6,"description":111,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":24},"01M6D3DRAK8VWESYDZDF8JJDSH","xray","Understand what just happened under the hood - learn AI by seeing it in action","https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fxray\u002FSKILL.md",[],[110,115,116],"learn","just",{"skillNo":118,"name":119,"nameZh":6,"description":120,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":121,"categories":122,"keywords":124,"publishedAt":24},"01M6D3DN7BZQDY6143A2QSDP9A","week-review","Review week's progress with concrete accomplishments (not fake percentages), pattern detection, and goal tracking.","https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fweek-review\u002FSKILL.md",[123],{"categoryNo":78,"categoryName":79,"taxonomyType":80,"taxonomyTypeDesc":81},[125,92,126,127,128,129,130],"week review","pattern detection","detection","tracking","goal","pattern",{"skillNo":132,"name":133,"nameZh":6,"description":134,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":135,"categories":136,"keywords":137,"publishedAt":24},"01M6D3DJ7F1TCFXBBSTQW82WE7","week-plan","Set weekly priorities and plan the week ahead with intelligent suggestions based on goals, calendar shape, and task effort.","https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fweek-plan\u002FSKILL.md",[],[138,21,139,140,141,142,143,144],"week plan","calendar","SET","shape","goals","task","priorities",{"skillNo":146,"name":147,"nameZh":6,"description":148,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":149,"categories":150,"keywords":151,"publishedAt":24},"01M6D3DF90D7RC8ZJE4AW2P9VC","triage","Strategically route orphaned files and extract scattered tasks","https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftriage\u002FSKILL.md",[],[147,152,153,154,155,156],"tasks","fast","extract","files","route",{"skillNo":158,"name":159,"nameZh":6,"description":160,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":161,"categories":162,"keywords":167,"publishedAt":24},"01M6D3DC9BBHK5TASE34SQYXSQ","trello-setup","Connect Trello to Amp for visual Kanban task sync","https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Ftrello-setup\u002FSKILL.md",[163,166],{"categoryNo":164,"categoryName":165,"taxonomyType":80,"taxonomyTypeDesc":81},"01JBSEEDCAT000000000000010","后端与接口",{"categoryNo":78,"categoryName":79,"taxonomyType":80,"taxonomyTypeDesc":81},[168,169,89,170,93,91,171,143,172,173,174,175,176],"trello","trello setup","sync","token","kanban","bun","visual","api","server"]