[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79H40S1Q22DRBSTH0SNF93J":3,"skill-same:01M79H40S1Q22DRBSTH0SNF93J":47},{"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":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":6,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M79H40S1Q22DRBSTH0SNF93J","experiment-design-matrix",null,"Use when planning experiments, ablations, or evaluation sweeps — before a costly run series, or when adding new cells (e.g. ablations) to a running campaign. Turns pre-registered hypotheses into a factor matrix with fair baselines, metrics, seeds, and stopping criteria. Not needed for trivial probes and smoke tests.","yibocat\u002Fprismnext","yibocat","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F98038316?v=4","01M79H3HDXFS2RDP1ED51TDF87","yibo","01M79H3HEG17B8ZH9AD0A06RBK",4,"MIT","https:\u002F\u002Fgithub.com\u002Fyibocat\u002Fprismnext\u002Fblob\u002Fmaster\u002Fresources\u002Fteams\u002Fprismnext.core\u002Fskills\u002Fexperiment-design-matrix\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: experiment-design-matrix\ndescription: Use when planning experiments, ablations, or evaluation sweeps — before a costly run series, or when adding new cells (e.g. ablations) to a running campaign. Turns pre-registered hypotheses into a factor matrix with fair baselines, metrics, seeds, and stopping criteria. Not needed for trivial probes and smoke tests.\nlicense: MIT\n---\n\n# Experiment Design Matrix\n\nDesign before looping runs. This skill receives the **pre-registered claims**\nfrom hypothesis design (via `.brief.md`) and expands them into a concrete\nfactor matrix — so every later run is interpretable against what was\npre-committed.\n\n**Proportionality rule**: the design record scales with the cost of the\nruns. A seconds-long probe or smoke test needs one sentence in chat; a sweep\nthat burns dozens of runs needs the full matrix below. Never write a design\ndoc heavier than the experiments it governs.\n\nOne design doc covers a whole **campaign** (a matrix of runs), not each\nindividual run — per-run records are kept automatically by the experiment\ntools.\n\n## When to use\n\n- Planning a new experiment, ablation study, or evaluation sweep\n- Adding cells (ablations, controls, new levels) to a campaign already running\n- Turning pre-registered hypotheses into a concrete factor matrix\n- The user is about to fire ad-hoc scripts with no design record\n\n## Workflow\n\n1. **Anchor** — `research-brief-read` to pick up the pre-registered claims,\n   metrics, and thresholds this matrix serves (optional for small probes,\n   never a gate).\n2. **Scaffold first** — `experiment-log` list → create → scaffold under the\n   workspace path; the design doc (`design.md`) is the island's first file.\n3. **Write the design doc** —\n   - Question being answered, linked to the claim it tests\n   - Factors × levels (independent variables)\n   - Baselines and controls — and why each comparison is *fair*\n   - Metrics with precise definitions\n   - Seeds \u002F repeats\n   - Stopping criteria\n   - Artifact paths that must survive\n4. **Cheap-first order** — smoke test → single matrix cell → full sweep.\n5. **Confirm cost, then execute** — before the full sweep, state the\n   estimate (runs × per-run time × resources) and confirm with the user via\n   `question`; smoke tests and single cells run without ceremony. Execute\n   with `experiment-run` (real artifact paths; batch runs by dependency;\n   summarize with ```artifact fences when helpful).\n6. **Close the loop** — `results-snapshot` for key results; killed cells\n   stay recorded in the matrix — never silently drop unfavorable runs.\n\n## Done when\n\n- Every factor has levels; every comparison has a fairness argument.\n- Stopping criteria are written before the sweep starts.\n- A smoke test has passed before any full sweep.\n- The design doc lives in the island, next to the runs it governs.\n\n## Boundaries\n\n- **Sample size, power, and test selection** → `statistical-rigor`; this\n  skill records *what* was decided, not the statistics behind it.\n- **Runtime\u002Fenvironment details** (shared `.prismnext\u002F.venv`, gates) live on\n  the experiment tools — follow them, do not restate them here.\n\n## Rules\n\n- Prefer fewer, interpretable runs over blind hyperparameter thrashing.\n- When an ablation cannot answer the claim, redesign — do not only add seeds.\n- Mid-campaign additions extend `design.md` in place; do not fork a second\n  design doc for the same campaign.\n","# Experiment Design Matrix\n\nDesign before looping runs. This skill receives the **pre-registered claims**\nfrom hypothesis design (via `.brief.md`) and expands them into a concrete\nfactor matrix — so every later run is interpretable against what was\npre-committed.\n\n**Proportionality rule**: the design record scales with the cost of the\nruns. A seconds-long probe or smoke test needs one sentence in chat; a sweep\nthat burns dozens of runs needs the full matrix below. Never write a design\ndoc heavier than the experiments it governs.\n\nOne design doc covers a whole **campaign** (a matrix of runs), not each\nindividual run — per-run records are kept automatically by the experiment\ntools.\n\n## When to use\n\n- Planning a new experiment, ablation study, or evaluation sweep\n- Adding cells (ablations, controls, new levels) to a campaign already running\n- Turning pre-registered hypotheses into a concrete factor matrix\n- The user is about to fire ad-hoc scripts with no design record\n\n## Workflow\n\n1. **Anchor** — `research-brief-read` to pick up the pre-registered claims,\n   metrics, and thresholds this matrix serves (optional for small probes,\n   never a gate).\n2. **Scaffold first** — `experiment-log` list → create → scaffold under the\n   workspace path; the design doc (`design.md`) is the island's first file.\n3. **Write the design doc** —\n   - Question being answered, linked to the claim it tests\n   - Factors × levels (independent variables)\n   - Baselines and controls — and why each comparison is *fair*\n   - Metrics with precise definitions\n   - Seeds \u002F repeats\n   - Stopping criteria\n   - Artifact paths that must survive\n4. **Cheap-first order** — smoke test → single matrix cell → full sweep.\n5. **Confirm cost, then execute** — before the full sweep, state the\n   estimate (runs × per-run time × resources) and confirm with the user via\n   `question`; smoke tests and single cells run without ceremony. Execute\n   with `experiment-run` (real artifact paths; batch runs by dependency;\n   summarize with ```artifact fences when helpful).\n6. **Close the loop** — `results-snapshot` for key results; killed cells\n   stay recorded in the matrix — never silently drop unfavorable runs.\n\n## Done when\n\n- Every factor has levels; every comparison has a fairness argument.\n- Stopping criteria are written before the sweep starts.\n- A smoke test has passed before any full sweep.\n- The design doc lives in the island, next to the runs it governs.\n\n## Boundaries\n\n- **Sample size, power, and test selection** → `statistical-rigor`; this\n  skill records *what* was decided, not the statistics behind it.\n- **Runtime\u002Fenvironment details** (shared `.prismnext\u002F.venv`, gates) live on\n  the experiment tools — follow them, do not restate them here.\n\n## Rules\n\n- Prefer fewer, interpretable runs over blind hyperparameter thrashing.\n- When an ablation cannot answer the claim, redesign — do not only add seeds.\n- Mid-campaign additions extend `design.md` in place; do not fork a second\n  design doc for the same campaign.\n",{"name":5,"license":15,"description":7},"master","bea8e76eec5d154324b095a78e731876fc136f75","resources\u002Fteams\u002Fprismnext.core\u002Fskills\u002Fexperiment-design-matrix",0,"https:\u002F\u002Fgithub.com\u002Fyibocat\u002Fprismnext","https:\u002F\u002Fprismnext.pages.dev\u002F","https:\u002F\u002Fgithub.com\u002Fyibocat","Data Science and Artificial Intelligence","2026-09-05 14:31:57",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add yibocat\u002Fprismnext --skill experiment-design-matrix",{"tool":39,"command":40},"bunx","bunx skills add yibocat\u002Fprismnext --skill experiment-design-matrix",{"tool":42,"command":43},"pnpm","pnpm dlx skills add yibocat\u002Fprismnext --skill experiment-design-matrix","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fyibocat\u002Fprismnext\u002Fblob\u002Fmaster\u002Fresources\u002Fteams\u002Fprismnext.core\u002Fskills\u002Fexperiment-design-matrix\u002FSKILL.md",true,[],[48,56,63,70,77,84],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":52,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":19},"01M79H88CRCDM6364CQV3AXTX8","demo-multi-file","Local test skill with an extra notes file","Apache-2.0","https:\u002F\u002Fgithub.com\u002Fyibocat\u002Fprismnext\u002Fblob\u002Fmaster\u002Ftests\u002Ffixtures\u002Fskill-registry\u002Fdemo-multi-file\u002FSKILL.md",[],[],{"skillNo":57,"name":58,"nameZh":6,"description":59,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":19},"01M79H84P8BEMKYP2EAJEJ12QQ","writing-results","Use when drafting a Results \u002F Experiments \u002F Analysis section — narrating verified evidence from run receipts, with complete statistical reporting, derivation analysis, and honest treatment of negative results. Every number traces to a run; every claim survives the audit.","https:\u002F\u002Fgithub.com\u002Fyibocat\u002Fprismnext\u002Fblob\u002Fmaster\u002Fresources\u002Fteams\u002Fprismnext.core\u002Fskills\u002Fwriting-results\u002FSKILL.md",[],[],{"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":69,"publishedAt":19},"01M79H80M7D6BJJ0NFJ74589E4","writing-related-work","Use when drafting a Related Work \u002F literature review section or a standalone synthesis note — organizing prior work into a grounded, cited narrative or comparing methods across papers, from the project library plus staged external discovery. Never cites from memory.","https:\u002F\u002Fgithub.com\u002Fyibocat\u002Fprismnext\u002Fblob\u002Fmaster\u002Fresources\u002Fteams\u002Fprismnext.core\u002Fskills\u002Fwriting-related-work\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M79H7WTP45W9KJQKQHX22PBR","writing-preliminaries","Use when drafting a Preliminaries \u002F Background \u002F Notation section — definitions, problem setup, and formal statements the rest of the paper builds on, with notation consistency and machine-checked formal claims. Reference patterns, not molds; include only what later sections actually use.","https:\u002F\u002Fgithub.com\u002Fyibocat\u002Fprismnext\u002Fblob\u002Fmaster\u002Fresources\u002Fteams\u002Fprismnext.core\u002Fskills\u002Fwriting-preliminaries\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":19},"01M79H7RYRFZ3J74Z15JNG6FD6","writing-methods","Use when drafting a Methods \u002F Methodology \u002F Approach \u002F Model section — expounding the ideas and the reasoning behind the method, grounded in the project's brief, notes, and prior discussions, with empirical details from run receipts. Structure follows the idea's shape; reference angles, not fixed templates.","https:\u002F\u002Fgithub.com\u002Fyibocat\u002Fprismnext\u002Fblob\u002Fmaster\u002Fresources\u002Fteams\u002Fprismnext.core\u002Fskills\u002Fwriting-methods\u002FSKILL.md",[],[],{"skillNo":85,"name":86,"nameZh":6,"description":87,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M79H7MPZB0KTWZBDFVN9NKP2","writing-introduction","Use when drafting or rewriting a paper's Introduction — framing the problem, positioning against prior work, and stating contributions as checkable claims, grounded in the brief, notes, and the library. Reference rhetorical patterns, adapted to the paper — never applied as molds.","https:\u002F\u002Fgithub.com\u002Fyibocat\u002Fprismnext\u002Fblob\u002Fmaster\u002Fresources\u002Fteams\u002Fprismnext.core\u002Fskills\u002Fwriting-introduction\u002FSKILL.md",[],[]]