[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6ERSKDDS8CT7FDPHBSW4F68":3,"skill-same:01M6ERSKDDS8CT7FDPHBSW4F68":62},{"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":28,"linkOnly":29,"markdownContent":30,"markdownBody":31,"frontMatter":32,"refName":33,"commitSha":34,"skillDirectory":35,"forksCount":36,"repositoryHtmlUrl":37,"repositoryHomepageUrl":6,"ownerHtmlUrl":38,"ownerBlogUrl":39,"ownerDescription":40,"updatedAt":41,"install":42,"resources":55},"01M6ERSKDDS8CT7FDPHBSW4F68","session-handoff",null,"Use when approaching context limits (75-85%) to write comprehensive handoff documents for session continuity. Also use when continuing from a handoff document.","nocksock\u002Fsuperpowers","nocksock","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F22164?v=4","01M6ERQWRTZARGFM7AX8HB89GZ","Nils Riedemann","01M6ERQWXFH5876Z6HEDH9TWYS",1,"MIT","https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers\u002Fblob\u002Fmain\u002Fskills\u002Fsession-handoff\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27],"handoff","session handoff","session","documents","write","context","use","document","continuity","2026-11-10 10:12:30",false,"---\nname: session-handoff\ndescription: \"Use when approaching context limits (75-85%) to write comprehensive handoff documents for session continuity. Also use when continuing from a handoff document.\"\n---\n\n# Session Handoff\n\n## Overview\n\nWrite comprehensive handoff documents when approaching context limits, and continue sessions from handoff documents.\n\n**Core principle:** The next session has ZERO context. Write for a stranger.\n\n**When to use:**\n- Context usage reaches 75-85% and work remains\n- User asks to pause and continue later\n- Continuing a session from a handoff document\n\n## CRITICAL: Understanding Context Loss\n\n**The new session will have ZERO context from the current session.**\n\nThis is not like \"pausing\" - this is a complete reset. The next agent:\n- Has no memory of this conversation\n- Does not know what files you looked at\n- Does not know what decisions you made or why\n- Does not know what the user said\n- Does not know what you learned about the codebase\n- Only knows what is written in the handoff document\n\n**If it's not in the handoff document, it does not exist for the next agent.**\n\nYou are not \"handing off to yourself\" - you are writing instructions for a stranger who happens to have the same capabilities. Everything you currently \"just know\" from this session MUST be explicitly written down or it is lost forever.\n\n## Monitoring Context Usage\n\nA hook injects context usage updates at 25%, 50%, 75%, then every 5% until 95%, then every 1% until 100%. You can also call the `checklimits` tool proactively.\n\nAt each context update, assess:\n- How much work remains?\n- Is there risk of running out before completing the task?\n- Early updates (25%, 50%) usually don't have enough info to judge\n\n## When to Wind Down\n\n**Target: Start winddown at 80-90% context usage.**\n\n- Not too early (waste context capacity)\n- Not too late (need context for winddown steps, and model performance degrades after 90-95%)\n\n**Trigger winddown when:**\n- Context usage is 75-85% AND\n- Remaining work likely exceeds remaining context AND\n- You can reach a reasonable stopping point\n\n## Writing a Handoff Document\n\n### Step 1: Finish Current Atomic Task\n\nComplete to a stopping point. Don't leave things half-done.\n\n### Step 2: Write Handoff Document\n\nYou MUST use the template in `.\u002Fhandoff-template.md`. \n\n**MUST NOT:**\n- Skip sections marked MUST\n- Write brief summaries assuming the next agent \"will figure it out\"\n- Use vague language like \"continue implementing\"\n\n**MUST:**\n- Fill every MUST section completely\n- Be specific about file paths, line numbers, exact state\n- Explain the \"why\" for every decision\n- Capture user preferences and quirks explicitly\n\n### Step 3: Save to Predictable Location\n\nSave to: `docs\u002Fsession-handoffs\u002FYYYY-MM-DD-HH-MM-\u003Ctopic>.md`\n\nCreate the directory if it doesn't exist.\n\n### Step 4: Ask User to Continue\n\nUse the question tool:\n\n> \"I'm approaching my context limit. I've written a detailed handoff document ready to continue in a new session. Ready to continue?\"\n> Options: Yes, continue in new session \u002F Not yet, I have questions\n\n### Step 5: Provide Continuation Prompt\n\n```\nI'm continuing a previous session. The handoff document is at:\ndocs\u002Fsession-handoffs\u002F[filename].md\n\nRead that file and continue where the previous session left off.\n```\n\nThe user will copy this into a new session.\n\n**For deep work sessions, use this prompt instead:**\n```\nI'm continuing a previous deep work session. Use the deep-work-session skill to continue. The handoff document is at:\ndocs\u002Fsession-handoffs\u002F[filename].md\n\nRead that file and continue the deep work session autonomously.\n```\n\n## Continuing From a Handoff Document\n\nWhen the user provides a handoff file path to continue from:\n\n### CRITICAL: You Have No Prior Context\n\n**You are not \"continuing\" - you are starting fresh with a briefing document.**\n\nYou have no memory of the previous session. You do not know:\n- What the user said before\n- What was tried and failed\n- What decisions were made or why\n- What the user's preferences are\n- What files are relevant\n\nThe handoff document is your ONLY source of context. You MUST read it completely and trust its contents.\n\n### Step 1: Read Handoff Document\n\nYou MUST read the file completely before taking any action. It contains:\n- Session preferences (if any were set)\n- Task context and progress\n- What was done, what remains\n- Technical decisions and reasoning\n- Open questions or blockers\n- User preferences and quirks\n\n### Step 2: Restore Session State\n\nYou MUST NOT re-ask preferences that are in the handoff document. Announce:\n\n```\nContinuing from [handoff file].\n\nRestored context:\n- [Key preference 1]\n- [Key preference 2]\n\nProgress: [X of Y tasks complete]\nResuming from: [current task]\n```\n\n### Step 3: Resume Work\n\n- Pick up exactly where the previous session left off\n- Follow the same preferences documented in the handoff\n- Continue with the same working style\n\n### Step 4: Handle Chained Handoffs\n\nIf this is a continuation of a continuation (chain of handoffs):\n- The handoff document should reference prior handoffs if relevant\n- Consolidate understanding, don't re-read entire chain\n- Focus on current state and remaining work\n\n## Red Flags\n\n**MUST NOT:**\n- Write a sparse handoff document - the next agent has ZERO context\n- Assume the next agent \"will figure it out\" - it will not, it cannot\n- Skip sections in the handoff template - every MUST section is required\n- Write vague next steps like \"continue implementing\" - be specific\n- Re-ask preferences when continuing from a handoff - trust the document\n- Ignore context usage warnings until it's too late to write a proper handoff\n\n**MUST:**\n- Write handoff documents as if explaining to a stranger (because you are)\n- Include the \"why\" for every technical decision, not just the \"what\"\n- Capture user preferences and quirks explicitly\n- Monitor context usage and plan winddown at 80-90%\n- Use `checklimits` tool proactively if unsure about remaining context\n\n## Integration\n\n**Used by:**\n- `deep-work-session` - for handling context limits in autonomous sessions\n- Any long-running session approaching context limits\n\n**Works with:**\n- All skills that establish session preferences (those should be captured in handoff)\n","# Session Handoff\n\n## Overview\n\nWrite comprehensive handoff documents when approaching context limits, and continue sessions from handoff documents.\n\n**Core principle:** The next session has ZERO context. Write for a stranger.\n\n**When to use:**\n- Context usage reaches 75-85% and work remains\n- User asks to pause and continue later\n- Continuing a session from a handoff document\n\n## CRITICAL: Understanding Context Loss\n\n**The new session will have ZERO context from the current session.**\n\nThis is not like \"pausing\" - this is a complete reset. The next agent:\n- Has no memory of this conversation\n- Does not know what files you looked at\n- Does not know what decisions you made or why\n- Does not know what the user said\n- Does not know what you learned about the codebase\n- Only knows what is written in the handoff document\n\n**If it's not in the handoff document, it does not exist for the next agent.**\n\nYou are not \"handing off to yourself\" - you are writing instructions for a stranger who happens to have the same capabilities. Everything you currently \"just know\" from this session MUST be explicitly written down or it is lost forever.\n\n## Monitoring Context Usage\n\nA hook injects context usage updates at 25%, 50%, 75%, then every 5% until 95%, then every 1% until 100%. You can also call the `checklimits` tool proactively.\n\nAt each context update, assess:\n- How much work remains?\n- Is there risk of running out before completing the task?\n- Early updates (25%, 50%) usually don't have enough info to judge\n\n## When to Wind Down\n\n**Target: Start winddown at 80-90% context usage.**\n\n- Not too early (waste context capacity)\n- Not too late (need context for winddown steps, and model performance degrades after 90-95%)\n\n**Trigger winddown when:**\n- Context usage is 75-85% AND\n- Remaining work likely exceeds remaining context AND\n- You can reach a reasonable stopping point\n\n## Writing a Handoff Document\n\n### Step 1: Finish Current Atomic Task\n\nComplete to a stopping point. Don't leave things half-done.\n\n### Step 2: Write Handoff Document\n\nYou MUST use the template in `.\u002Fhandoff-template.md`. \n\n**MUST NOT:**\n- Skip sections marked MUST\n- Write brief summaries assuming the next agent \"will figure it out\"\n- Use vague language like \"continue implementing\"\n\n**MUST:**\n- Fill every MUST section completely\n- Be specific about file paths, line numbers, exact state\n- Explain the \"why\" for every decision\n- Capture user preferences and quirks explicitly\n\n### Step 3: Save to Predictable Location\n\nSave to: `docs\u002Fsession-handoffs\u002FYYYY-MM-DD-HH-MM-\u003Ctopic>.md`\n\nCreate the directory if it doesn't exist.\n\n### Step 4: Ask User to Continue\n\nUse the question tool:\n\n> \"I'm approaching my context limit. I've written a detailed handoff document ready to continue in a new session. Ready to continue?\"\n> Options: Yes, continue in new session \u002F Not yet, I have questions\n\n### Step 5: Provide Continuation Prompt\n\n```\nI'm continuing a previous session. The handoff document is at:\ndocs\u002Fsession-handoffs\u002F[filename].md\n\nRead that file and continue where the previous session left off.\n```\n\nThe user will copy this into a new session.\n\n**For deep work sessions, use this prompt instead:**\n```\nI'm continuing a previous deep work session. Use the deep-work-session skill to continue. The handoff document is at:\ndocs\u002Fsession-handoffs\u002F[filename].md\n\nRead that file and continue the deep work session autonomously.\n```\n\n## Continuing From a Handoff Document\n\nWhen the user provides a handoff file path to continue from:\n\n### CRITICAL: You Have No Prior Context\n\n**You are not \"continuing\" - you are starting fresh with a briefing document.**\n\nYou have no memory of the previous session. You do not know:\n- What the user said before\n- What was tried and failed\n- What decisions were made or why\n- What the user's preferences are\n- What files are relevant\n\nThe handoff document is your ONLY source of context. You MUST read it completely and trust its contents.\n\n### Step 1: Read Handoff Document\n\nYou MUST read the file completely before taking any action. It contains:\n- Session preferences (if any were set)\n- Task context and progress\n- What was done, what remains\n- Technical decisions and reasoning\n- Open questions or blockers\n- User preferences and quirks\n\n### Step 2: Restore Session State\n\nYou MUST NOT re-ask preferences that are in the handoff document. Announce:\n\n```\nContinuing from [handoff file].\n\nRestored context:\n- [Key preference 1]\n- [Key preference 2]\n\nProgress: [X of Y tasks complete]\nResuming from: [current task]\n```\n\n### Step 3: Resume Work\n\n- Pick up exactly where the previous session left off\n- Follow the same preferences documented in the handoff\n- Continue with the same working style\n\n### Step 4: Handle Chained Handoffs\n\nIf this is a continuation of a continuation (chain of handoffs):\n- The handoff document should reference prior handoffs if relevant\n- Consolidate understanding, don't re-read entire chain\n- Focus on current state and remaining work\n\n## Red Flags\n\n**MUST NOT:**\n- Write a sparse handoff document - the next agent has ZERO context\n- Assume the next agent \"will figure it out\" - it will not, it cannot\n- Skip sections in the handoff template - every MUST section is required\n- Write vague next steps like \"continue implementing\" - be specific\n- Re-ask preferences when continuing from a handoff - trust the document\n- Ignore context usage warnings until it's too late to write a proper handoff\n\n**MUST:**\n- Write handoff documents as if explaining to a stranger (because you are)\n- Include the \"why\" for every technical decision, not just the \"what\"\n- Capture user preferences and quirks explicitly\n- Monitor context usage and plan winddown at 80-90%\n- Use `checklimits` tool proactively if unsure about remaining context\n\n## Integration\n\n**Used by:**\n- `deep-work-session` - for handling context limits in autonomous sessions\n- Any long-running session approaching context limits\n\n**Works with:**\n- All skills that establish session preferences (those should be captured in handoff)\n",{"name":5,"description":7},"main","52d2d5a29297631b5e3ff61fbc8ca3c6505a26e2","skills\u002Fsession-handoff",0,"https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers","https:\u002F\u002Fgithub.com\u002Fnocksock","http:\u002F\u002Fbleepbloop.studio\u002F","Hi, I'm Nils. I build interfaces for the web.","2026-03-19 04:55:50",{"source":8,"skillName":5,"commands":43,"prompt":53,"zipAvailable":54},[44,47,50],{"tool":45,"command":46},"npx","npx skills add nocksock\u002Fsuperpowers --skill session-handoff",{"tool":48,"command":49},"bunx","bunx skills add nocksock\u002Fsuperpowers --skill session-handoff",{"tool":51,"command":52},"pnpm","pnpm dlx skills add nocksock\u002Fsuperpowers --skill session-handoff","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers\u002Fblob\u002Fmain\u002Fskills\u002Fsession-handoff\u002FSKILL.md",true,[56],{"path":57,"resourceType":58,"resourceTypeDesc":59,"sizeBytes":60,"sourceHtmlUrl":61},"handoff-template.md","REFERENCE","参考文档",6985,"https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers\u002Fblob\u002Fmain\u002Fskills\u002Fsession-handoff\u002Fhandoff-template.md",[63,88,104,120,133,144],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":82,"publishedAt":28},"01M6ERV5R58PKM81EH15XEYPMF","writing-skills","Use when creating new skills, editing existing skills, or verifying skills work before deployment","https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-skills\u002FSKILL.md",[69,74,77],{"categoryNo":70,"categoryName":71,"taxonomyType":72,"taxonomyTypeDesc":73},"01JBSEEDCAT000000000000008","测试与质量","SCENARIO","使用场景",{"categoryNo":75,"categoryName":76,"taxonomyType":72,"taxonomyTypeDesc":73},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":78,"categoryName":79,"taxonomyType":80,"taxonomyTypeDesc":81},"01JBSEEDCAT000000000000036","内容创作者","PROFESSION","适用职业",[83,84,25,85,86,87],"writing skills","writing","deployment","work","editing",{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":96,"publishedAt":28},"01M6ERV2JX5DJV2EHSC6FXHEHN","writing-plans","Use when you have a spec or requirements for a multi-step task, before touching code","https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-plans\u002FSKILL.md",[94,95],{"categoryNo":75,"categoryName":76,"taxonomyType":72,"taxonomyTypeDesc":73},{"categoryNo":78,"categoryName":79,"taxonomyType":80,"taxonomyTypeDesc":81},[97,98,84,99,100,25,101,102,103],"plans","writing plans","spec","code","task","requirements","multi",{"skillNo":105,"name":106,"nameZh":6,"description":107,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":108,"categories":109,"keywords":115,"publishedAt":28},"01M6ERTZS70NCVS375DTD8HGKE","writing good code","use this skill whenever you write code. it provides a process for creating good, composable apis.","https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-good-code\u002FSKILL.md",[110,111,114],{"categoryNo":75,"categoryName":76,"taxonomyType":72,"taxonomyTypeDesc":73},{"categoryNo":112,"categoryName":113,"taxonomyType":80,"taxonomyTypeDesc":81},"01JBSEEDCAT000000000000030","开发工程师",{"categoryNo":78,"categoryName":79,"taxonomyType":80,"taxonomyTypeDesc":81},[106,116,100,84,117,25,118,23,119],"good","it","process","composable",{"skillNo":121,"name":122,"nameZh":6,"description":123,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":124,"categories":125,"keywords":126,"publishedAt":28},"01M6ERTVVRPKZVWK6M8EXPB5QE","write-review-elixir","Review Elixir code. ALWAYS use this skill before or after you added code written in elixir to ensure high code quality. Also use use this when asked to review existing elixir code.","https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers\u002Fblob\u002Fmain\u002Fskills\u002Fwrite-or-review-elixir\u002FSKILL.md",[],[127,23,128,129,130,25,131,100,132],"write review elixir","elixir","review","elixir code","quality","code quality",{"skillNo":134,"name":135,"nameZh":6,"description":136,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":137,"categories":138,"keywords":139,"publishedAt":28},"01M6ERTRAM8PSJ9JTAW5J4ABFD","whats-next","Answers what to do next, what to work on next, where we're at","https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers\u002Fblob\u002Fmain\u002Fskills\u002Fwhats-next\u002FSKILL.md",[],[140,141,142,86,143],"next","whats next","work on","do",{"skillNo":145,"name":146,"nameZh":6,"description":147,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":148,"categories":149,"keywords":150,"publishedAt":28},"01M6ERTNG9NT8SQH5FHN91JRTS","verification-before-completion","Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always","https:\u002F\u002Fgithub.com\u002Fnocksock\u002Fsuperpowers\u002Fblob\u002Fmain\u002Fskills\u002Fverification-before-completion\u002FSKILL.md",[],[151,152,153,154,155,25,86,156,157,158],"verification before completion","verification","commands","claim","evidence","complete","output","about"]