[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78DZ1D77XG8EXM2943672F6":3,"skill-same:01M78DZ1D77XG8EXM2943672F6":51},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":28,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M78DZ1D77XG8EXM2943672F6","deviation-log",null,"Deviation log — log departures from the plan as work hits unknowns, any domain. Use on \"implementation notes\", \"record a deviation\", \"where did we diverge from the plan?\".","ajitta\u002Fknow-your-unknowns","ajitta","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F10125909?v=4","01M78DYNVE4A35YFEBKBJNKJX4","01M78DYNWK9J1VH59EKN8D93A9",2,"MIT","https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns\u002Fblob\u002Fmain\u002Fplugins\u002Funknowns\u002Fskills\u002Fdeviation-log\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: deviation-log\ndescription: >\n  Deviation log — log departures from the plan as work hits unknowns, any domain. Use on\n  \"implementation notes\", \"record a deviation\", \"where did we diverge from the plan?\".\n---\n\n# Deviation Log\n\nThis skill serves what neither the user nor you know yet, and surfaces both sides' unknowns.\n\nWhen you and the user both seem to understand the request, do it without this method: if its result carries weight (it ships, is sent or is hard to undo), first read back in 2–3 lines the request as you understand it, the assumptions you are making and what outcome counts as done, as the whole reply, and do it once the user confirms; a self-evident request (fix a typo) needs no read-back. When the read-back exposes a mismatch, route by who does not know: the user → the **teach-me** skill; you → the **interview** skill; neither of you → the **blindspot** skill. A misread weighty request costs the work built on it; a read-back costs one exchange.\n\nWhen work hits a situation the plan or spec does not cover (an unknown neither side saw), **pick the conservative option, log it, and keep going**. The log is where the user sees where the territory departed from the map, and why.\n\n## Procedure\n\n1. Take the entry from the request: text given with the skill name is the **Situation found**, and the rest of the entry is filled from context. With no text, log the most recent deviation of this session that meets the Logging Criteria below and is not yet in the deviation log, or say there is none.\n2. Append the entry (a one-line **Discovery** when no decision is needed yet; see Entry Format) to the deviation log before ending the turn, creating the file if it is missing:\n\n   | Work | Log file | Heading when created |\n   |---|---|---|\n   | Code | `IMPLEMENTATION_NOTES.md` at the project root | `# Implementation Notes — plan deviation log` |\n   | Anything else | the file the user names, or `DEVIATIONS.md` beside the deliverable | `# Deviation log — \u003Cdeliverable>` |\n\n   The **buy-in** and **quiz** skills read that log. Without a tool, do the step in the conversation and say what could not be checked.\n3. While working, check for unlogged deviations (as the Logging Criteria define them) at each checkpoint (a commit, a shared draft) and before handing the work over, and append each one as in step 2.\n4. When a deviation invalidates a decision in the plan document, mark that decision in the plan as superseded by the entry's date, so the plan stays the map the work follows.\n\n## Entry Format\n\nDeviation entry — two blocks that do **not** carry equal weight. Fill Observed from the artifacts first, and keep it as found when you write Attributed:\n\n```markdown\n## [YYYY-MM-DD] \u003Ctask\u002Ffeature name>\n\n**Observed** — the record, auditable against the work\n- **Situation found**: what unplanned thing was hit\n- **Deviation from plan**: what the plan said (quote it, or \"not covered\")\n- **Response chosen**: what the work now does\n\n**Attributed** — hypothesis, not evidence\n- **Reason for choice**: why that approach\n- **Alternatives considered**: discarded options and why discarded\n- **Risk\u002Ffollow-up check**: if this decision is wrong, where it shows up, and what only the user can judge\n```\n\n**Why the split.** A stated reason is a hypothesis: models and people both produce fluent rationales that omit what actually drove the choice. Observed earns trust only by being checkable against the work as it stands (the diff, the document, the booking); Attributed is a claim to test later. On conflict, the work as it stands wins.\n\nWhen context does not say which alternatives were weighed, fill **Alternatives considered** with the plausible options you can name, each with why it loses here, marked *(inferred, not stated)*. If no other option was viable, say why. A bare \"none\" or \"not recorded\" leaves the field unfilled. Attributed is hypothesis already, so a marked inference stays honest, and a reviewer can challenge the choice only against named alternatives.\n\nWhen the choice rests on something only the user can know (their priorities, a person's preference, a local rule), name it in **Risk\u002Ffollow-up check**: the conservative option was your guess, and the wrap-up puts it to the user.\n\nA **Discovery** is the lighter entry, one line in the same file: the territory differs from what the plan assumed and no decision is needed yet. Log what was assumed \u002F what is true.\n\n## Logging Criteria\n\n- **Log**: decisions affecting design, behavior or compatibility; points where the spec's interpretation diverges; an existing structure (code, a venue, a dataset) differing from what the plan assumed; workarounds; new dependencies, suppliers or commitments.\n- **Don't log**: trivial wording or syntax fixes, formatting, names, self-evident details, and an item the plan said to check that came back as planned, since it departs from nothing and a value such as a password does not belong in a shared log.\n\n## Wrap-up\n\nBefore ending the session or handing the work over (a PR, a submitted draft, a handoff):\n\n1. List first the entries whose decision touches the structure of the work, what others see or were promised (e.g. a UX\u002FAPI contract, a published date), data or money, or security and safety: the user reviews those calls before anything else, starting with the ones that rest on something only the user can know.\n2. Summarize the remaining entries.\n3. Write the **fold back into the plan** block: 3 copyable bullets on what this changes about attempt #2, so the next run does not rediscover today's surprises.\n4. When you largely produced the work being handed over, suggest continuing with the **quiz** skill, so the user can explain it, these deviations included, before the hand-over.\n5. If the work needs approval, carry this log's recorded deviations and unresolved risks into the **buy-in** skill doc as \"known limitations\".\n","# Deviation Log\n\nThis skill serves what neither the user nor you know yet, and surfaces both sides' unknowns.\n\nWhen you and the user both seem to understand the request, do it without this method: if its result carries weight (it ships, is sent or is hard to undo), first read back in 2–3 lines the request as you understand it, the assumptions you are making and what outcome counts as done, as the whole reply, and do it once the user confirms; a self-evident request (fix a typo) needs no read-back. When the read-back exposes a mismatch, route by who does not know: the user → the **teach-me** skill; you → the **interview** skill; neither of you → the **blindspot** skill. A misread weighty request costs the work built on it; a read-back costs one exchange.\n\nWhen work hits a situation the plan or spec does not cover (an unknown neither side saw), **pick the conservative option, log it, and keep going**. The log is where the user sees where the territory departed from the map, and why.\n\n## Procedure\n\n1. Take the entry from the request: text given with the skill name is the **Situation found**, and the rest of the entry is filled from context. With no text, log the most recent deviation of this session that meets the Logging Criteria below and is not yet in the deviation log, or say there is none.\n2. Append the entry (a one-line **Discovery** when no decision is needed yet; see Entry Format) to the deviation log before ending the turn, creating the file if it is missing:\n\n   | Work | Log file | Heading when created |\n   |---|---|---|\n   | Code | `IMPLEMENTATION_NOTES.md` at the project root | `# Implementation Notes — plan deviation log` |\n   | Anything else | the file the user names, or `DEVIATIONS.md` beside the deliverable | `# Deviation log — \u003Cdeliverable>` |\n\n   The **buy-in** and **quiz** skills read that log. Without a tool, do the step in the conversation and say what could not be checked.\n3. While working, check for unlogged deviations (as the Logging Criteria define them) at each checkpoint (a commit, a shared draft) and before handing the work over, and append each one as in step 2.\n4. When a deviation invalidates a decision in the plan document, mark that decision in the plan as superseded by the entry's date, so the plan stays the map the work follows.\n\n## Entry Format\n\nDeviation entry — two blocks that do **not** carry equal weight. Fill Observed from the artifacts first, and keep it as found when you write Attributed:\n\n```markdown\n## [YYYY-MM-DD] \u003Ctask\u002Ffeature name>\n\n**Observed** — the record, auditable against the work\n- **Situation found**: what unplanned thing was hit\n- **Deviation from plan**: what the plan said (quote it, or \"not covered\")\n- **Response chosen**: what the work now does\n\n**Attributed** — hypothesis, not evidence\n- **Reason for choice**: why that approach\n- **Alternatives considered**: discarded options and why discarded\n- **Risk\u002Ffollow-up check**: if this decision is wrong, where it shows up, and what only the user can judge\n```\n\n**Why the split.** A stated reason is a hypothesis: models and people both produce fluent rationales that omit what actually drove the choice. Observed earns trust only by being checkable against the work as it stands (the diff, the document, the booking); Attributed is a claim to test later. On conflict, the work as it stands wins.\n\nWhen context does not say which alternatives were weighed, fill **Alternatives considered** with the plausible options you can name, each with why it loses here, marked *(inferred, not stated)*. If no other option was viable, say why. A bare \"none\" or \"not recorded\" leaves the field unfilled. Attributed is hypothesis already, so a marked inference stays honest, and a reviewer can challenge the choice only against named alternatives.\n\nWhen the choice rests on something only the user can know (their priorities, a person's preference, a local rule), name it in **Risk\u002Ffollow-up check**: the conservative option was your guess, and the wrap-up puts it to the user.\n\nA **Discovery** is the lighter entry, one line in the same file: the territory differs from what the plan assumed and no decision is needed yet. Log what was assumed \u002F what is true.\n\n## Logging Criteria\n\n- **Log**: decisions affecting design, behavior or compatibility; points where the spec's interpretation diverges; an existing structure (code, a venue, a dataset) differing from what the plan assumed; workarounds; new dependencies, suppliers or commitments.\n- **Don't log**: trivial wording or syntax fixes, formatting, names, self-evident details, and an item the plan said to check that came back as planned, since it departs from nothing and a value such as a password does not belong in a shared log.\n\n## Wrap-up\n\nBefore ending the session or handing the work over (a PR, a submitted draft, a handoff):\n\n1. List first the entries whose decision touches the structure of the work, what others see or were promised (e.g. a UX\u002FAPI contract, a published date), data or money, or security and safety: the user reviews those calls before anything else, starting with the ones that rest on something only the user can know.\n2. Summarize the remaining entries.\n3. Write the **fold back into the plan** block: 3 copyable bullets on what this changes about attempt #2, so the next run does not rediscover today's surprises.\n4. When you largely produced the work being handed over, suggest continuing with the **quiz** skill, so the user can explain it, these deviations included, before the hand-over.\n5. If the work needs approval, carry this log's recorded deviations and unresolved risks into the **buy-in** skill doc as \"known limitations\".\n",{"name":5,"description":7},"main","f1a43ab1ad56e66493708bb8383ab8cd7dbda389","plugins\u002Funknowns\u002Fskills\u002Fdeviation-log",0,"https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns","https:\u002F\u002Fajitta.github.io\u002Fknow-your-unknowns\u002F","https:\u002F\u002Fgithub.com\u002Fajitta","2026-11-13 00:17:49",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add ajitta\u002Fknow-your-unknowns --skill deviation-log",{"tool":37,"command":38},"bunx","bunx skills add ajitta\u002Fknow-your-unknowns --skill deviation-log",{"tool":40,"command":41},"pnpm","pnpm dlx skills add ajitta\u002Fknow-your-unknowns --skill deviation-log","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns\u002Fblob\u002Fmain\u002Fplugins\u002Funknowns\u002Fskills\u002Fdeviation-log\u002FSKILL.md",true,[45],{"path":46,"resourceType":47,"resourceTypeDesc":48,"sizeBytes":49,"sourceHtmlUrl":50},"agents\u002Fopenai.yaml","REFERENCE","参考文档",245,"https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns\u002Fblob\u002Fmain\u002Fplugins\u002Funknowns\u002Fskills\u002Fdeviation-log\u002Fagents\u002Fopenai.yaml",[52,59,66,73,80,87],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":18},"01M78E0W9N76BZGDDG14X7EAFX","unknowns-loop","Run the know-your-unknowns workflow end to end, ending with an outcome check. Use on \"unknowns loop\", \"run the operating loop\", \"know your unknowns\". A bare \"loop\" is NOT this.","https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns\u002Fblob\u002Fmain\u002Fskills\u002Funknowns-loop\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":18},"01M78E0SQ91CY0RCXX51KK7T55","tweakable-plan","The tweakable plan — decisions most likely to change come first, with alternatives, for any work. Use on \"plan this\", \"make a plan\", \"tweakable plan\".","https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns\u002Fblob\u002Fmain\u002Fskills\u002Ftweakable-plan\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":18},"01M78E0PV11V4HF079XHEDBR48","teach-me","Teach me my unknowns — an explainer for a field's vocabulary and mental model, so vague requests become precise. Use on \"teach me\", \"make me an explainer\". Not a plan investigation.","https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns\u002Fblob\u002Fmain\u002Fskills\u002Fteach-me\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":18},"01M78E0KZXF9H786W6AF590Y27","reference","Point at a reference — read an example, show your reading for correction before building on it; runs even when the adaptation looks simple. Use on \"use this as a reference\", \"make it like this\".","https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns\u002Fblob\u002Fmain\u002Fskills\u002Freference\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":18},"01M78E0GWRA5C2WZJS4D6DYGBF","quiz","Comprehension quiz — explain the work, then quiz you so you can defend it at handover; runs even when a quiz is easy to write. Use on \"quiz me\", \"test my understanding\", or before merging.","https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns\u002Fblob\u002Fmain\u002Fskills\u002Fquiz\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":18},"01M78E0DSWFA5FV09TNEVVNR8T","prototypes","Prototypes — several versions of anything to react to, each on a different philosophy. Use on \"divergent prototypes\", \"design options\", \"versions to react to\", \"know it when I see it\".","https:\u002F\u002Fgithub.com\u002Fajitta\u002Fknow-your-unknowns\u002Fblob\u002Fmain\u002Fskills\u002Fprototypes\u002FSKILL.md",[],[]]