[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7CZKNPHDEWBKK3TS9FVZ1B5":3,"skill-same:01M7CZKNPHDEWBKK3TS9FVZ1B5":46},{"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":30,"updatedAt":31,"install":32,"resources":45},"01M7CZKNPHDEWBKK3TS9FVZ1B5","good-map",null,"What makes a good map of a Lean proof (the target proved from helper lemmas stated with sorry) — read it before you write or judge a map.","Infini-AI-Lab\u002Fhermit","Infini-AI-Lab","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F149595684?v=4","01M7CZJ2B2FWAKCN12205RA5YT","01M7CZJ2BNJ2V36H0YNVDF7MD9",522,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab\u002Fhermit\u002Fblob\u002Fmain\u002Fhermit\u002Fadapter\u002Ftemplates\u002Flean_collab\u002Fskills\u002Fgood-map\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: good-map\ndescription: What makes a good map of a Lean proof (the target proved from helper lemmas stated with sorry) — read it before you write or judge a map.\n---\n\n# A good map\n\nA map is what the whole team inherits: once it is registered, five agents spend\ntheir time on its parts. A weak map costs more than a late one. A split of a part into\nsub-parts is a map of that part, and the same rules apply, with the part in place\nof the target.\n\n**Choose the route by what its parts will cost.** When there are several routes, count\nthe unproven pieces each one needs and how hard each piece is — a route whose parts\nneed only crude estimates beats one that needs sharp ones — then commit; do not keep\nreopening the choice. Prefer a route the library already supports.\n\n**Every part is smaller than the target.** A part that is the target in other\nwords — the same claim with a definition unfolded, a constant rewritten, or a\nhypothesis moved — splits nothing, even if the map has several parts. A map whose one\npart holds the whole difficulty is not a map yet. Ask of each part: is it strictly\neasier than the target, and why?\n\n**Split until each part has a proof sketch.** For every part, ask: can I say in a\nfew sentences how it is proved? If not, it is still too coarse: prove it from\nsmaller `sorry` lemmas, which become its children — the map gets another level\nthere. If yes, stop: those sentences go in its docstring as the evidence. If the\nsketch has cases, the cases are the parts. Splitting lowers the difficulty; a part\nsplit into pieces each as hard as the part lowers nothing.\n\n**The hard part is named, not hidden.** Know where the difficulty of the proof\nlies and give it its own part (or parts). A map made of one small simplification\nplus \"everything else\" is not a decomposition. Whoever writes the map starts on its\nhardest part at once, in its simplest statement.\n\n**Every part is believed true, with evidence.** For each part, have a reason it\nholds: a numerical check over a range, a known theorem, a short argument. Put the\nreason in the part's docstring. Check the edge cases (zero, small values, the\nboundary of a range) — a false part sends the team down a dead end. The evidence\nis a numerical check, a known theorem or a short argument — not a Lean proof.\n\n**Statements are exact.** The hypotheses each part needs are stated; types and\ncasts are the ones the proof will use; the part says precisely what the glue\nuses. Settle the definitions before you hand out the parts: a definition that changes\nlater invalidates the work done on it.\n\n**The parts fit together simply.** The target's proof from the parts (the glue)\nshould be short and obviously right. If the glue itself needs a long argument,\nthat argument is a missing part.\n\n**Each part is a unit one agent can own.** Not chopped into dozens of trivial\nsteps. Machinery several parts share (a general lemma, a definition with its basic\nfacts) is its own part, stated in general form, with the specific object replaced by\na variable.\n\n**It uses what the team has.** Library lemmas that state a step close it at once;\nuse them as parts where they fit. Build on the findings on the message board, and\nsay whose idea the map follows.\n","# A good map\n\nA map is what the whole team inherits: once it is registered, five agents spend\ntheir time on its parts. A weak map costs more than a late one. A split of a part into\nsub-parts is a map of that part, and the same rules apply, with the part in place\nof the target.\n\n**Choose the route by what its parts will cost.** When there are several routes, count\nthe unproven pieces each one needs and how hard each piece is — a route whose parts\nneed only crude estimates beats one that needs sharp ones — then commit; do not keep\nreopening the choice. Prefer a route the library already supports.\n\n**Every part is smaller than the target.** A part that is the target in other\nwords — the same claim with a definition unfolded, a constant rewritten, or a\nhypothesis moved — splits nothing, even if the map has several parts. A map whose one\npart holds the whole difficulty is not a map yet. Ask of each part: is it strictly\neasier than the target, and why?\n\n**Split until each part has a proof sketch.** For every part, ask: can I say in a\nfew sentences how it is proved? If not, it is still too coarse: prove it from\nsmaller `sorry` lemmas, which become its children — the map gets another level\nthere. If yes, stop: those sentences go in its docstring as the evidence. If the\nsketch has cases, the cases are the parts. Splitting lowers the difficulty; a part\nsplit into pieces each as hard as the part lowers nothing.\n\n**The hard part is named, not hidden.** Know where the difficulty of the proof\nlies and give it its own part (or parts). A map made of one small simplification\nplus \"everything else\" is not a decomposition. Whoever writes the map starts on its\nhardest part at once, in its simplest statement.\n\n**Every part is believed true, with evidence.** For each part, have a reason it\nholds: a numerical check over a range, a known theorem, a short argument. Put the\nreason in the part's docstring. Check the edge cases (zero, small values, the\nboundary of a range) — a false part sends the team down a dead end. The evidence\nis a numerical check, a known theorem or a short argument — not a Lean proof.\n\n**Statements are exact.** The hypotheses each part needs are stated; types and\ncasts are the ones the proof will use; the part says precisely what the glue\nuses. Settle the definitions before you hand out the parts: a definition that changes\nlater invalidates the work done on it.\n\n**The parts fit together simply.** The target's proof from the parts (the glue)\nshould be short and obviously right. If the glue itself needs a long argument,\nthat argument is a missing part.\n\n**Each part is a unit one agent can own.** Not chopped into dozens of trivial\nsteps. Machinery several parts share (a general lemma, a definition with its basic\nfacts) is its own part, stated in general form, with the specific object replaced by\na variable.\n\n**It uses what the team has.** Library lemmas that state a step close it at once;\nuse them as parts where they fit. Build on the findings on the message board, and\nsay whose idea the map follows.\n",{"name":5,"description":7},"main","e81353897a1b5174178fe9fdf537f68862cc18ee","hermit\u002Fadapter\u002Ftemplates\u002Flean_collab\u002Fskills\u002Fgood-map",46,"https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab\u002Fhermit","https:\u002F\u002Finfini-ai-lab.github.io\u002Fhermit\u002F","https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab","Next Generation AI algorithms and systems","2026-11-13 10:07:23",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add Infini-AI-Lab\u002Fhermit --skill good-map",{"tool":38,"command":39},"bunx","bunx skills add Infini-AI-Lab\u002Fhermit --skill good-map",{"tool":41,"command":42},"pnpm","pnpm dlx skills add Infini-AI-Lab\u002Fhermit --skill good-map","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab\u002Fhermit\u002Fblob\u002Fmain\u002Fhermit\u002Fadapter\u002Ftemplates\u002Flean_collab\u002Fskills\u002Fgood-map\u002FSKILL.md",true,[],[47,53,59,65,71,77],{"skillNo":48,"name":49,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":18},"01M7CZMQVA5C6BBSRD2AVFAGE5","keep-journal","https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab\u002Fhermit\u002Fblob\u002Fmain\u002Fhermit\u002Fagent\u002Fevolver\u002Fmethods\u002Fmwm\u002Fmaterials\u002Fskills\u002Fkeep-journal\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":6,"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},"01M7CZMK8QC1FXQP17QNAAG94S","failure-discipline","https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab\u002Fhermit\u002Fblob\u002Fmain\u002Fhermit\u002Fagent\u002Fevolver\u002Fmethods\u002Fmwm\u002Fmaterials\u002Fskills\u002Ffailure-discipline\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":18},"01M7CZMFXPFX20FV51NP1GHY6N","design-representation","https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab\u002Fhermit\u002Fblob\u002Fmain\u002Fhermit\u002Fagent\u002Fevolver\u002Fmethods\u002Fmwm\u002Fmaterials\u002Fskills\u002Fdesign-representation\u002FSKILL.md",[],[],{"skillNo":66,"name":67,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":18},"01M7CZMCKY8AA74JVACE7WQ2ZN","decide-executable","https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab\u002Fhermit\u002Fblob\u002Fmain\u002Fhermit\u002Fagent\u002Fevolver\u002Fmethods\u002Fmwm\u002Fmaterials\u002Fskills\u002Fdecide-executable\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":18},"01M7CZMARG8WRCWPKTY6EB6MNB","curate-brief","https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab\u002Fhermit\u002Fblob\u002Fmain\u002Fhermit\u002Fagent\u002Fevolver\u002Fmethods\u002Fmwm\u002Fmaterials\u002Fskills\u002Fcurate-brief\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":18},"01M7CZM6TBNYYVTZZNF0DCZK55","create-subagent","https:\u002F\u002Fgithub.com\u002FInfini-AI-Lab\u002Fhermit\u002Fblob\u002Fmain\u002Fhermit\u002Fagent\u002Fevolver\u002Fmethods\u002Fmwm\u002Fmaterials\u002Fskills\u002Fcreate-subagent\u002FSKILL.md",[],[]]