[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6D350GJBK0DPXFH4321P2Q1":3,"skill-same:01M6D350GJBK0DPXFH4321P2Q1":51},{"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":23,"linkOnly":24,"markdownContent":25,"markdownBody":26,"frontMatter":27,"refName":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":34,"ownerDescription":35,"updatedAt":36,"install":37,"resources":50},"01M6D350GJBK0DPXFH4321P2Q1","amp-whats-new",null,"Check for system improvements (learnings + Claude updates)","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-whats-new\u002FSKILL.md",[],[19,20,21,22],"amp whats new","system","check","learnings","2026-11-10 10:12:30",false,"---\nname: amp-whats-new\ndescription: Check for system improvements (learnings + Claude updates)\n---\n\nCheck for improvements to your Amp system from TWO sources:\n1. **Learnings from your usage** - Mistakes, patterns, opportunities captured during sessions\n2. **New Claude Code capabilities** - Features that could enhance your workflows\n\n## Usage\n\n```\n\u002Famp-whats-new              # Review learnings + check for Claude updates\n\u002Famp-whats-new --full       # Include detailed explanations\n\u002Famp-whats-new --learnings  # Only review session learnings\n\u002Famp-whats-new --claude     # Only check Claude Code updates\n```\n\n## Arguments\n\n$MODE: Optional flags\n- `--full` - Detailed explanations of each feature\n- `--learnings` - Only review learnings (skip Claude check)\n- `--claude` - Only check Claude updates (skip learnings)\n\n---\n\n## Process\n\n### Step 0: Review Session Learnings (Unless --claude flag)\n\n**Check for learnings to review:**\n\n1. Read files in `System\u002FSession_Learnings\u002F` from last 30 days\n2. Read `06-Resources\u002FLearnings\u002FMistake_Patterns.md` - check for new patterns\n3. Read `06-Resources\u002FLearnings\u002FWorking_Preferences.md` - check for trends\n\n**Extract improvement opportunities:**\n- Recurring mistakes → suggest preventive measures\n- Gaps in documentation → suggest additions\n- Workflow inefficiencies → suggest automations\n- User patterns → suggest customizations\n\n**Present findings:**\n\n```\n? LEARNINGS FROM YOUR USAGE\n\nSince last review: [X] session learnings captured\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? PATTERNS IDENTIFIED\n\n1. [Pattern name]\n   Observed: [frequency] times\n   Impact: [what this affects]\n   Suggestion: [concrete improvement]\n   \n2. [Pattern name]\n   ...\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? SUGGESTED IMPROVEMENTS\n\nBased on what you've learned, here's how to improve Amp:\n\n1. [Improvement name]\n   Problem: [What's not working smoothly]\n   Solution: [Specific change to make]\n   Files: [What to update]\n   Effort: Low\u002FMedium\u002FHigh\n   \n2. [Improvement name]\n   ...\n\nWant me to implement any of these? (Enter number)\n```\n\n**If no learnings:**\nSkip this section or show: \"No session learnings captured yet. The system will learn as you use it.\"\n\n---\n\n### Step 1: Read Current State (Unless --learnings flag)\n\nLoad `System\u002Fclaude-code-state.json`:\n\n```json\n{\n  \"last_check\": \"2026-01-15\",\n  \"last_version_seen\": \"1.0.28\",\n  \"features_seen\": [\"hooks\", \"sub-agents\", \"skills\", \"commands\", \"mcp\"]\n}\n```\n\nIf file missing or `last_check` is null, treat as first run.\n\n### Step 2: Fetch Current Changelog\n\nUse `WebSearch` to find the latest Claude Code changelog:\n- Search: \"Claude Code changelog 2026\" or \"Anthropic Claude Code releases\"\n- Primary source: Anthropic's official documentation\n- Fallback: GitHub releases, official blog posts\n\nFocus on:\n- New features and capabilities\n- Breaking changes\n- Deprecations\n- Performance improvements\n\n### Step 3: Compare and Surface Changes\n\nIdentify what's new since `last_check`:\n\n**For each new feature:**\n1. What it does (plain English, 1-2 sentences)\n2. Why it matters for PKM users\n3. How you could use it in Amp (concrete example)\n4. Effort to adopt (Low\u002FMedium\u002FHigh)\n\n### Step 4: Present Claude Findings\n\n**If Claude Code updates found:**\n\n```\n? CLAUDE CODE UPDATES\n\nLast checked: [date] (X days ago)\nCurrent version: [version]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? NEW FEATURES\n\n1. [Feature Name]\n   What: [Plain English description]\n   For you: [How this could improve Amp]\n   Effort: Low\n\n2. [Feature Name]\n   What: [Description]\n   For you: [Specific improvement idea]\n   Effort: Medium\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? SUGGESTED IMPROVEMENTS\n\nBased on what's new, here are concrete things you could add to Amp:\n\n1. [Improvement name]\n   Uses: [Which new feature]\n   What it does: [Specific description]\n   Pillar: [Which pillar it supports]\n\nWant me to implement any of these? (Enter number)\nOr run `\u002Famp-improve` to workshop custom ideas.\n```\n\n**If no Claude updates:**\n\n```\n✅ Claude Code is up to date!\n\nLast checked: Today\nCurrent version: [version]\n\nNo new features since your last check.\n```\n\n---\n\n### Combined View (Default)\n\nWhen running without flags, show BOTH sections:\n\n```\n? AMP SYSTEM IMPROVEMENT REVIEW\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? FROM YOUR USAGE\n[Session learnings section]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? FROM CLAUDE CODE\n[Claude updates section]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? COMBINED IMPROVEMENTS\n\nPulling together learnings + new capabilities:\n\n1. [Improvement combining both]\n   Uses: [Your pattern] + [New Claude feature]\n   Impact: [What this unlocks]\n   \nWant me to implement any of these? (Enter number or 'all')\n```\n\n### Step 5: Update State\n\n**CRITICAL:** After presenting findings, you MUST update `System\u002Fclaude-code-state.json` using the Write tool.\n\n**Steps:**\n1. Read current state file to preserve existing `features_seen` array\n2. Add any newly discovered features to the array (avoid duplicates)\n3. Update `last_check` to today's date (YYYY-MM-DD format)\n4. Update `last_version_seen` if a new version was found\n5. Write the updated JSON back to `System\u002Fclaude-code-state.json`\n\n**Example update:**\n```json\n{\n  \"last_check\": \"2026-01-28\",\n  \"last_version_seen\": \"1.0.32\",\n  \"features_seen\": [\n    \"hooks\",\n    \"sub-agents\",\n    \"skills\", \n    \"commands\",\n    \"mcp\",\n    \"parallel-sub-agents\",\n    \"async-hooks\"\n  ]\n}\n```\n\n**Verification:** After writing, confirm to user: \"State file updated - next check will compare against today's findings.\"\n\n---\n\n### Step 6: Backlog Synthesis (NEW — Innovation Concierge)\n\nAfter presenting findings to the user, automatically create or enrich backlog ideas:\n\n1. **Call `synthesize_changelog()`** from Improvements MCP\n   - This scans the changelog for Amp-relevant features\n   - Creates new AI-authored ideas for novel capabilities\n   - Enriches existing ideas with \"Why Now?\" evidence when a platform feature strengthens them\n\n2. **Call `synthesize_learnings()`** from Improvements MCP\n   - Scans pending session learnings for improvement opportunities\n   - Creates ideas from learnings that have concrete \"suggested fix\" entries\n   - Enriches existing ideas when learnings relate to known backlog items\n\n3. **Report synthesis results to user:**\n\n```\n? BACKLOG SYNTHESIS\n\nChangelog: Scanned X entries → Created Y new ideas, enriched Z existing\nLearnings: Scanned X entries → Created Y new ideas, enriched Z existing\n\nTop ideas created\u002Fenriched:\n1. [idea-XXX] Title (action: created\u002Fenriched)\n2. [idea-XXX] Title (action: created\u002Fenriched)\n3. [idea-XXX] Title (action: created\u002Fenriched)\n\nRun `\u002Famp-backlog` to see full ranked backlog.\n```\n\n**This step is what connects external intelligence to your improvement backlog.**\nWithout it, findings are presented and forgotten. With it, every relevant discovery\nbecomes a tracked, ranked improvement opportunity.\n\n---\n\n## Full Mode (--full)\n\nWhen `--full` is provided, include educational deep-dives:\n\nFor each feature, add:\n\n```\n? DEEP DIVE: [Feature Name]\n\n**What it is:**\n[2-3 paragraph explanation of the capability]\n\n**How it works:**\n[Technical explanation with examples]\n\n**Real-world example:**\n[Concrete scenario showing the feature in action]\n\n**In Amp, you could:**\n- [Specific application 1]\n- [Specific application 2]\n\n**To implement:**\n1. [Step 1]\n2. [Step 2]\n3. [Step 3]\n```\n\n---\n\n## Feature Categories\n\nWhen evaluating relevance, categorize features:\n\n| Category | Relevance to Amp | Examples |\n|----------|------------------|----------|\n| **Automation** | High | Hooks, triggers, scheduled tasks |\n| **Performance** | Medium | Faster models, caching |\n| **Context** | High | Memory, skills, knowledge bases |\n| **Integration** | High | MCP improvements, new protocols |\n| **UI\u002FUX** | Low | IDE features, visual changes |\n| **Developer** | Low | API changes, SDK updates |\n\nFocus on High relevance categories. Mention Medium. Skip Low unless asked.\n\n---\n\n## Capability Reference\n\nCurrent Claude Code features to track:\n\n| Feature | What It Does | Amp Potential |\n|---------|--------------|---------------|\n| **Commands** | User-triggered workflows | `\u002Fplan`, `\u002Fdaily-review`, etc. |\n| **Skills** | Reusable behaviors, always loaded | Person lookup, writing style |\n| **Hooks** | Auto-triggers at events | Session start, file changes |\n| **Sub-agents** | Parallel workers, isolated context | Research, analysis |\n| **MCP** | External service connections | Calendar, tasks, email |\n| **Memory** | Cross-session persistence | Preferences, learnings |\n| **Tools** | Built-in capabilities | File ops, search, terminal |\n\n---\n\n## Example Output\n\n```\n? Claude Code Updates\n\nLast checked: 2026-01-15 (8 days ago)\nCurrent version: 1.0.32\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? NEW FEATURES\n\n1. Async Hooks\n   What: Hooks can now run without blocking the main conversation\n   For you: Faster session starts — load context in background\n   Effort: Low\n\n2. Sub-agent Communication\n   What: Sub-agents can now pass data back to parent\n   For you: Research agent could update your notes directly\n   Effort: Medium\n\n3. MCP Resource Subscriptions\n   What: MCP servers can push updates, not just respond to queries\n   For you: Get notified when calendar changes, tasks update\n   Effort: High (requires MCP server changes)\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? SUGGESTED IMPROVEMENTS\n\n1. Background Context Loading\n   Uses: Async Hooks\n   What: Load your learnings and today's plan while you type\n   Pillar: Productivity\n\n2. Auto-Research on New Topics\n   Uses: Sub-agent Communication\n   What: When you mention a new company, auto-research in background\n   Pillar: Deal Support\n\nWant me to implement any of these? (Enter 1 or 2)\n```\n\n---\n\n## Error Handling\n\n**If web search fails:**\n> \"Couldn't fetch the latest changelog. Here's what I know was current as of [last_check]:\n> [List known capabilities]\n> \n> Try again later, or check manually at docs.anthropic.com\"\n\n**If state file is corrupted:**\n> \"Your capability state file has an issue. I'll create a fresh one and do a full scan.\"\n\nThen proceed with first-run behavior.\n\n---\n\n## Behaviors\n\n### Always Do\n- Keep explanations in plain English, not developer jargon\n- Tie every feature back to concrete Amp improvements\n- Update the state file after every check\n- Offer to implement suggestions\n\n### Never Do\n- List features without explaining relevance\n- Skip the state update\n- Overwhelm with every minor change (focus on impactful features)\n- Assume user knows Claude Code internals\n\n---\n\n## Related Commands\n\n- `\u002Famp-improve` — Full design partner (includes this + workshopping + audit)\n- `\u002Fcreate-mcp` — Build new integrations when new MCP features enable them\n\n## Track Usage (Silent)\n\nUpdate `System\u002Fusage_log.md` to mark what's new check as used.\n\n**Analytics (Silent):**\n\nCall `track_event` with event_name `whats_new_viewed` and properties:\n- `update_available` (boolean)\n\nThis only fires if the user has opted into analytics. No action needed if it returns \"analytics_disabled\".\n","Check for improvements to your Amp system from TWO sources:\n1. **Learnings from your usage** - Mistakes, patterns, opportunities captured during sessions\n2. **New Claude Code capabilities** - Features that could enhance your workflows\n\n## Usage\n\n```\n\u002Famp-whats-new              # Review learnings + check for Claude updates\n\u002Famp-whats-new --full       # Include detailed explanations\n\u002Famp-whats-new --learnings  # Only review session learnings\n\u002Famp-whats-new --claude     # Only check Claude Code updates\n```\n\n## Arguments\n\n$MODE: Optional flags\n- `--full` - Detailed explanations of each feature\n- `--learnings` - Only review learnings (skip Claude check)\n- `--claude` - Only check Claude updates (skip learnings)\n\n---\n\n## Process\n\n### Step 0: Review Session Learnings (Unless --claude flag)\n\n**Check for learnings to review:**\n\n1. Read files in `System\u002FSession_Learnings\u002F` from last 30 days\n2. Read `06-Resources\u002FLearnings\u002FMistake_Patterns.md` - check for new patterns\n3. Read `06-Resources\u002FLearnings\u002FWorking_Preferences.md` - check for trends\n\n**Extract improvement opportunities:**\n- Recurring mistakes → suggest preventive measures\n- Gaps in documentation → suggest additions\n- Workflow inefficiencies → suggest automations\n- User patterns → suggest customizations\n\n**Present findings:**\n\n```\n? LEARNINGS FROM YOUR USAGE\n\nSince last review: [X] session learnings captured\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? PATTERNS IDENTIFIED\n\n1. [Pattern name]\n   Observed: [frequency] times\n   Impact: [what this affects]\n   Suggestion: [concrete improvement]\n   \n2. [Pattern name]\n   ...\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? SUGGESTED IMPROVEMENTS\n\nBased on what you've learned, here's how to improve Amp:\n\n1. [Improvement name]\n   Problem: [What's not working smoothly]\n   Solution: [Specific change to make]\n   Files: [What to update]\n   Effort: Low\u002FMedium\u002FHigh\n   \n2. [Improvement name]\n   ...\n\nWant me to implement any of these? (Enter number)\n```\n\n**If no learnings:**\nSkip this section or show: \"No session learnings captured yet. The system will learn as you use it.\"\n\n---\n\n### Step 1: Read Current State (Unless --learnings flag)\n\nLoad `System\u002Fclaude-code-state.json`:\n\n```json\n{\n  \"last_check\": \"2026-01-15\",\n  \"last_version_seen\": \"1.0.28\",\n  \"features_seen\": [\"hooks\", \"sub-agents\", \"skills\", \"commands\", \"mcp\"]\n}\n```\n\nIf file missing or `last_check` is null, treat as first run.\n\n### Step 2: Fetch Current Changelog\n\nUse `WebSearch` to find the latest Claude Code changelog:\n- Search: \"Claude Code changelog 2026\" or \"Anthropic Claude Code releases\"\n- Primary source: Anthropic's official documentation\n- Fallback: GitHub releases, official blog posts\n\nFocus on:\n- New features and capabilities\n- Breaking changes\n- Deprecations\n- Performance improvements\n\n### Step 3: Compare and Surface Changes\n\nIdentify what's new since `last_check`:\n\n**For each new feature:**\n1. What it does (plain English, 1-2 sentences)\n2. Why it matters for PKM users\n3. How you could use it in Amp (concrete example)\n4. Effort to adopt (Low\u002FMedium\u002FHigh)\n\n### Step 4: Present Claude Findings\n\n**If Claude Code updates found:**\n\n```\n? CLAUDE CODE UPDATES\n\nLast checked: [date] (X days ago)\nCurrent version: [version]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? NEW FEATURES\n\n1. [Feature Name]\n   What: [Plain English description]\n   For you: [How this could improve Amp]\n   Effort: Low\n\n2. [Feature Name]\n   What: [Description]\n   For you: [Specific improvement idea]\n   Effort: Medium\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? SUGGESTED IMPROVEMENTS\n\nBased on what's new, here are concrete things you could add to Amp:\n\n1. [Improvement name]\n   Uses: [Which new feature]\n   What it does: [Specific description]\n   Pillar: [Which pillar it supports]\n\nWant me to implement any of these? (Enter number)\nOr run `\u002Famp-improve` to workshop custom ideas.\n```\n\n**If no Claude updates:**\n\n```\n✅ Claude Code is up to date!\n\nLast checked: Today\nCurrent version: [version]\n\nNo new features since your last check.\n```\n\n---\n\n### Combined View (Default)\n\nWhen running without flags, show BOTH sections:\n\n```\n? AMP SYSTEM IMPROVEMENT REVIEW\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? FROM YOUR USAGE\n[Session learnings section]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? FROM CLAUDE CODE\n[Claude updates section]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? COMBINED IMPROVEMENTS\n\nPulling together learnings + new capabilities:\n\n1. [Improvement combining both]\n   Uses: [Your pattern] + [New Claude feature]\n   Impact: [What this unlocks]\n   \nWant me to implement any of these? (Enter number or 'all')\n```\n\n### Step 5: Update State\n\n**CRITICAL:** After presenting findings, you MUST update `System\u002Fclaude-code-state.json` using the Write tool.\n\n**Steps:**\n1. Read current state file to preserve existing `features_seen` array\n2. Add any newly discovered features to the array (avoid duplicates)\n3. Update `last_check` to today's date (YYYY-MM-DD format)\n4. Update `last_version_seen` if a new version was found\n5. Write the updated JSON back to `System\u002Fclaude-code-state.json`\n\n**Example update:**\n```json\n{\n  \"last_check\": \"2026-01-28\",\n  \"last_version_seen\": \"1.0.32\",\n  \"features_seen\": [\n    \"hooks\",\n    \"sub-agents\",\n    \"skills\", \n    \"commands\",\n    \"mcp\",\n    \"parallel-sub-agents\",\n    \"async-hooks\"\n  ]\n}\n```\n\n**Verification:** After writing, confirm to user: \"State file updated - next check will compare against today's findings.\"\n\n---\n\n### Step 6: Backlog Synthesis (NEW — Innovation Concierge)\n\nAfter presenting findings to the user, automatically create or enrich backlog ideas:\n\n1. **Call `synthesize_changelog()`** from Improvements MCP\n   - This scans the changelog for Amp-relevant features\n   - Creates new AI-authored ideas for novel capabilities\n   - Enriches existing ideas with \"Why Now?\" evidence when a platform feature strengthens them\n\n2. **Call `synthesize_learnings()`** from Improvements MCP\n   - Scans pending session learnings for improvement opportunities\n   - Creates ideas from learnings that have concrete \"suggested fix\" entries\n   - Enriches existing ideas when learnings relate to known backlog items\n\n3. **Report synthesis results to user:**\n\n```\n? BACKLOG SYNTHESIS\n\nChangelog: Scanned X entries → Created Y new ideas, enriched Z existing\nLearnings: Scanned X entries → Created Y new ideas, enriched Z existing\n\nTop ideas created\u002Fenriched:\n1. [idea-XXX] Title (action: created\u002Fenriched)\n2. [idea-XXX] Title (action: created\u002Fenriched)\n3. [idea-XXX] Title (action: created\u002Fenriched)\n\nRun `\u002Famp-backlog` to see full ranked backlog.\n```\n\n**This step is what connects external intelligence to your improvement backlog.**\nWithout it, findings are presented and forgotten. With it, every relevant discovery\nbecomes a tracked, ranked improvement opportunity.\n\n---\n\n## Full Mode (--full)\n\nWhen `--full` is provided, include educational deep-dives:\n\nFor each feature, add:\n\n```\n? DEEP DIVE: [Feature Name]\n\n**What it is:**\n[2-3 paragraph explanation of the capability]\n\n**How it works:**\n[Technical explanation with examples]\n\n**Real-world example:**\n[Concrete scenario showing the feature in action]\n\n**In Amp, you could:**\n- [Specific application 1]\n- [Specific application 2]\n\n**To implement:**\n1. [Step 1]\n2. [Step 2]\n3. [Step 3]\n```\n\n---\n\n## Feature Categories\n\nWhen evaluating relevance, categorize features:\n\n| Category | Relevance to Amp | Examples |\n|----------|------------------|----------|\n| **Automation** | High | Hooks, triggers, scheduled tasks |\n| **Performance** | Medium | Faster models, caching |\n| **Context** | High | Memory, skills, knowledge bases |\n| **Integration** | High | MCP improvements, new protocols |\n| **UI\u002FUX** | Low | IDE features, visual changes |\n| **Developer** | Low | API changes, SDK updates |\n\nFocus on High relevance categories. Mention Medium. Skip Low unless asked.\n\n---\n\n## Capability Reference\n\nCurrent Claude Code features to track:\n\n| Feature | What It Does | Amp Potential |\n|---------|--------------|---------------|\n| **Commands** | User-triggered workflows | `\u002Fplan`, `\u002Fdaily-review`, etc. |\n| **Skills** | Reusable behaviors, always loaded | Person lookup, writing style |\n| **Hooks** | Auto-triggers at events | Session start, file changes |\n| **Sub-agents** | Parallel workers, isolated context | Research, analysis |\n| **MCP** | External service connections | Calendar, tasks, email |\n| **Memory** | Cross-session persistence | Preferences, learnings |\n| **Tools** | Built-in capabilities | File ops, search, terminal |\n\n---\n\n## Example Output\n\n```\n? Claude Code Updates\n\nLast checked: 2026-01-15 (8 days ago)\nCurrent version: 1.0.32\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? NEW FEATURES\n\n1. Async Hooks\n   What: Hooks can now run without blocking the main conversation\n   For you: Faster session starts — load context in background\n   Effort: Low\n\n2. Sub-agent Communication\n   What: Sub-agents can now pass data back to parent\n   For you: Research agent could update your notes directly\n   Effort: Medium\n\n3. MCP Resource Subscriptions\n   What: MCP servers can push updates, not just respond to queries\n   For you: Get notified when calendar changes, tasks update\n   Effort: High (requires MCP server changes)\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n? SUGGESTED IMPROVEMENTS\n\n1. Background Context Loading\n   Uses: Async Hooks\n   What: Load your learnings and today's plan while you type\n   Pillar: Productivity\n\n2. Auto-Research on New Topics\n   Uses: Sub-agent Communication\n   What: When you mention a new company, auto-research in background\n   Pillar: Deal Support\n\nWant me to implement any of these? (Enter 1 or 2)\n```\n\n---\n\n## Error Handling\n\n**If web search fails:**\n> \"Couldn't fetch the latest changelog. Here's what I know was current as of [last_check]:\n> [List known capabilities]\n> \n> Try again later, or check manually at docs.anthropic.com\"\n\n**If state file is corrupted:**\n> \"Your capability state file has an issue. I'll create a fresh one and do a full scan.\"\n\nThen proceed with first-run behavior.\n\n---\n\n## Behaviors\n\n### Always Do\n- Keep explanations in plain English, not developer jargon\n- Tie every feature back to concrete Amp improvements\n- Update the state file after every check\n- Offer to implement suggestions\n\n### Never Do\n- List features without explaining relevance\n- Skip the state update\n- Overwhelm with every minor change (focus on impactful features)\n- Assume user knows Claude Code internals\n\n---\n\n## Related Commands\n\n- `\u002Famp-improve` — Full design partner (includes this + workshopping + audit)\n- `\u002Fcreate-mcp` — Build new integrations when new MCP features enable them\n\n## Track Usage (Silent)\n\nUpdate `System\u002Fusage_log.md` to mark what's new check as used.\n\n**Analytics (Silent):**\n\nCall `track_event` with event_name `whats_new_viewed` and properties:\n- `update_available` (boolean)\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-whats-new",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":38,"prompt":48,"zipAvailable":49},[39,42,45],{"tool":40,"command":41},"npx","npx skills add hemory\u002Famp-legacy-do-not-use --skill amp-whats-new",{"tool":43,"command":44},"bunx","bunx skills add hemory\u002Famp-legacy-do-not-use --skill amp-whats-new",{"tool":46,"command":47},"pnpm","pnpm dlx skills add hemory\u002Famp-legacy-do-not-use --skill amp-whats-new","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fhemory\u002Famp-legacy-do-not-use\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Famp-whats-new\u002FSKILL.md",true,[],[52,89,98,112,127,139],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":68,"publishedAt":23},"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",[58,63],{"categoryNo":59,"categoryName":60,"taxonomyType":61,"taxonomyTypeDesc":62},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",{"categoryNo":64,"categoryName":65,"taxonomyType":66,"taxonomyTypeDesc":67},"01JBSEEDCAT000000000000034","行政与办公人员","PROFESSION","适用职业",[69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88],"zoom setup","setup","schedule","MCP","review","connect","transcript","search","stats","oauth2","confirmation","context","meetings","process","recordings","surfaces","scheduling","granola","time","distribution",{"skillNo":90,"name":91,"nameZh":6,"description":92,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":93,"categories":94,"keywords":95,"publishedAt":23},"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",[],[91,96,97],"learn","just",{"skillNo":99,"name":100,"nameZh":6,"description":101,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":102,"categories":103,"keywords":105,"publishedAt":23},"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",[104],{"categoryNo":59,"categoryName":60,"taxonomyType":61,"taxonomyTypeDesc":62},[106,73,107,108,109,110,111],"week review","pattern detection","detection","tracking","goal","pattern",{"skillNo":113,"name":114,"nameZh":6,"description":115,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":116,"categories":117,"keywords":118,"publishedAt":23},"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",[],[119,120,121,122,123,124,125,126],"week plan","plan","calendar","SET","shape","goals","task","priorities",{"skillNo":128,"name":129,"nameZh":6,"description":130,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":131,"categories":132,"keywords":133,"publishedAt":23},"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",[],[129,134,135,136,137,138],"tasks","fast","extract","files","route",{"skillNo":140,"name":141,"nameZh":6,"description":142,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":143,"categories":144,"keywords":149,"publishedAt":23},"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",[145,148],{"categoryNo":146,"categoryName":147,"taxonomyType":61,"taxonomyTypeDesc":62},"01JBSEEDCAT000000000000010","后端与接口",{"categoryNo":59,"categoryName":60,"taxonomyType":61,"taxonomyTypeDesc":62},[150,151,70,152,74,72,153,125,154,155,156,157,158],"trello","trello setup","sync","token","kanban","bun","visual","api","server"]