[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78D01HWFGS7QQ9DSEF0VB13":3,"skill-same:01M78D01HWFGS7QQ9DSEF0VB13":66},{"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":30,"commitSha":31,"skillDirectory":32,"forksCount":33,"repositoryHtmlUrl":34,"repositoryHomepageUrl":35,"ownerHtmlUrl":36,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":37,"install":38,"resources":51},"01M78D01HWFGS7QQ9DSEF0VB13","workspace-plans",null,"Use this skill when writing, revising, or executing a workspace plan — the plan\u002FREADME.md + phase-N-\u003Cslug>.md format the dashboard ingests. Don't use it for ad-hoc todo lists or in-repo design docs.","atretyak1985\u002Fswarmery","atretyak1985","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F1436737?v=4","01M78CTSX57YH0EYY9YGH0TWKX","Andrii Tretiak","01M78CTSXRPJPYQ7GNPMP45TYX",8,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Fplugins\u002Fcore\u002Fskills\u002Fworkspace-plans\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: workspace-plans\nversion: \"1.2.0\"\nowner: \"swarmery-core\"\ndescription: \"Use this skill when writing, revising, or executing a workspace plan — the plan\u002FREADME.md + phase-N-\u003Cslug>.md format the dashboard ingests. Don't use it for ad-hoc todo lists or in-repo design docs.\"\ncolor: cyan\ndocs:\n  status: draft\n  updated: 2026-10-21\n---\n\n# Purpose\n\nThe private workspace plan format is a contract: the control plane parses it\nto render plan progress, phase summaries, and spec coverage. A plan that\ndeviates from the format ships work the operator cannot see.\n\n# The contract (never violate)\n\n- Plans live at `\u003Cworkspace>\u002F\u003Cproject>\u002Fworkspace\u002Fworking\u002F{YYYY}\u002F{MM}\u002F{DD}\u002F{slug}\u002Fplan\u002F`;\n  task-id `yyyy-mm-dd-slug` is derived from the path — never encode the date\n  in the folder name.\n- `plan\u002FREADME.md` — an `## In plain words` section under the H1, objective,\n  architecture decisions with real file paths, the sequencing table with\n  exactly these header cells: `| # | Phase | Doc | Depends on |`, risks,\n  Definition of Done.\n- `plan\u002Fphase-N-\u003Cslug>.md` per phase (flat; `step-NN-*.md` is legacy\n  read-compat only) — each opening with an `## In plain words` section (after\n  the header rows, before `## Goal`), with a self-contained copy-paste executor prompt\n  whose Verify block is file-backed (a repo script or test by absolute path;\n  no heredoc, `python3 -` or `node -e` — see `resources\u002Fplan-format.md`),\n  `- [ ]` acceptance criteria checkable by command or boolean inspection, and\n  an empty `## Completion Report` section as the LAST section.\n- A `## Forecast` section in each phase doc, immediately before that stub,\n  holding one `kind: prior` yaml block (see `resources\u002Fplan-format.md`).\n  It is a prediction, not a limit: do whatever the phase actually needs.\n- `plan\u002Fmanifest.json` — machine-readable phase DAG (must pass\n  `python3 -m json.tool`).\n- Optional `plan\u002Fspec.md` with `- [ ] **SC-n** — \u003Ccriterion>` lines; then\n  every phase doc carries a `**Covers:** SC-…` header line (coverage is\n  linted).\n\nCriteria closed by a push, PR or merge start with `[LAND]` and name no branch\n(`- [ ] [LAND] phase PR merged into main`) — the header's `**Branch:**` is\ninformational and the daemon lands only the run branch `swarm\u002Fphase-\u003Cid>`.\nCriteria only a human can perform (production, a console, a manual check)\nstart with `[MANUAL]`. Neither is the executor's to tick or perform: the\ndaemon settles a run against the rest, refuses to start a `[MANUAL]`-only\nphase and lists it in Needs you, lists a phase with only `[LAND]` left as\nawaiting merge, and ticks `[LAND]` itself when the pull request for the run\nbranch merges — whether landed through the Review tab or opened outside the\nplatform and found by the landing poll or \"Find PR\". When the PR comes from\nany other branch, its `[LAND]` criteria are ticked by hand. The header rows\n`**Verify:**` (off | normal | strict), `**Review:**` (on | off) and the required `**Effort:**` (low | medium | high | xhigh)\nsit directly after `**Repo:** …`; L\u002FXL phases default to `normal` \u002F `on`. Details: `resources\u002Fplan-format.md` → \"Criterion classes\"\nand \"Header rows\".\n\nCriteria that lean on the engine follow five rules in `resources\u002Fplan-format.md`:\nspike a guarantee under every permission mode, live-check engine-boundary phases\nbefore `[LAND]`, append live-check evidence immediately, cite\n`resources\u002Fconfig-channels.md` for any settings-file criterion, and number\nmigrations at landing.\n\n# Resources\n\n- `resources\u002Fplan-format.md` — the full field-by-field format, examples, the\n  manifest schema, and the engine-facing criterion rules.\n- `resources\u002Fblast-radius.md` — the planner's blast-radius step: who calls\n  what a phase changes, and the `Blast radius:` line its `## Risks` opens with.\n- `resources\u002Fconfig-channels.md` — which settings tier Claude Code reads for\n  `pluginConfigs`, plugin `env` and `${VAR}`, with the CLI version measured.\n- `templates\u002Fspike-matrix.md` (plugin root) — the permission mode × surface\n  matrix for a spike.\n\n# Executor duties\n\nAfter reading the code the phase touches and before your first edit, add a\n`kind: posterior` block beside the prior in `## Forecast` — the\nprediction-not-a-limit rule above applies unchanged — and if the work then\ndiverges from it, say how and why in a short \"Where reality diverged\"\nparagraph of the Completion Report.\n\nTick each satisfied criterion `- [ ]` → `- [x]` immediately after verifying\nit — progress is derived only from these checkboxes. When a phase's last\nEXECUTABLE criterion is ticked (every one that is not `[LAND]`\u002F`[MANUAL]` —\nthose close later, without an executor), fill that doc's `## Completion Report` (what shipped,\nfiles\u002Fcommits, verification output, deviations, plus three mandatory fields —\n**Blocked calls**, **Delegation cost**, and the one-sentence **What would have\nmade this cheaper**; ≤50 lines) — the dashboard renders exactly that heading.\nEach of the three is written out even when empty (\"none\" \u002F \"Nothing, it was\nalready minimal\"); `resources\u002Fplan-format.md` carries their exact wording and\nis the source of truth. Never tick unsatisfied criteria; never archive\nwith unmet criteria or a missing SUMMARY.md.\n\nFull field-by-field format, examples, and the manifest schema:\n`resources\u002Fplan-format.md`.\n\n# How to use\n\n## What it does\n\nCarries the exact workspace-plan format the dashboard ingests: README sequencing table, phase docs with file-backed-verification executor prompts and checkbox criteria, manifest DAG, optional spec with SC coverage, and the tick\u002FCompletion Report duties for executors — including the report's Blocked calls, Delegation cost and \"what would have made this cheaper\" fields.\n\n## When to use it\n\nLoad it whenever you write a new plan, revise one, or execute one — any time plan files under a task dir's `plan\u002F` are created or updated.\n\n## How to invoke\n\nThe `@core:planner` and `@core:implementation-agent` agents list it in their `skills:`; load it manually with `\u002Fworkspace-plans` context or by reading `resources\u002Fplan-format.md` when editing plan files by hand.\n\n## Worked example\n\nA planner writing `plan\u002Fphase-2-api.md` includes: `**Covers:** SC-3`, a prompt an executor can run cold whose Verify block names `scripts\u002Fcheck-api.sh` rather than inlining a heredoc, three `- [ ]` criteria each with a check command, and ends the doc with an empty `## Completion Report` section. The executor later ticks the boxes one by one and fills the report — Blocked calls, Delegation cost and the one-sentence cheaper field included — and the dashboard's phase Summary tab shows exactly that section.\n","# Purpose\n\nThe private workspace plan format is a contract: the control plane parses it\nto render plan progress, phase summaries, and spec coverage. A plan that\ndeviates from the format ships work the operator cannot see.\n\n# The contract (never violate)\n\n- Plans live at `\u003Cworkspace>\u002F\u003Cproject>\u002Fworkspace\u002Fworking\u002F{YYYY}\u002F{MM}\u002F{DD}\u002F{slug}\u002Fplan\u002F`;\n  task-id `yyyy-mm-dd-slug` is derived from the path — never encode the date\n  in the folder name.\n- `plan\u002FREADME.md` — an `## In plain words` section under the H1, objective,\n  architecture decisions with real file paths, the sequencing table with\n  exactly these header cells: `| # | Phase | Doc | Depends on |`, risks,\n  Definition of Done.\n- `plan\u002Fphase-N-\u003Cslug>.md` per phase (flat; `step-NN-*.md` is legacy\n  read-compat only) — each opening with an `## In plain words` section (after\n  the header rows, before `## Goal`), with a self-contained copy-paste executor prompt\n  whose Verify block is file-backed (a repo script or test by absolute path;\n  no heredoc, `python3 -` or `node -e` — see `resources\u002Fplan-format.md`),\n  `- [ ]` acceptance criteria checkable by command or boolean inspection, and\n  an empty `## Completion Report` section as the LAST section.\n- A `## Forecast` section in each phase doc, immediately before that stub,\n  holding one `kind: prior` yaml block (see `resources\u002Fplan-format.md`).\n  It is a prediction, not a limit: do whatever the phase actually needs.\n- `plan\u002Fmanifest.json` — machine-readable phase DAG (must pass\n  `python3 -m json.tool`).\n- Optional `plan\u002Fspec.md` with `- [ ] **SC-n** — \u003Ccriterion>` lines; then\n  every phase doc carries a `**Covers:** SC-…` header line (coverage is\n  linted).\n\nCriteria closed by a push, PR or merge start with `[LAND]` and name no branch\n(`- [ ] [LAND] phase PR merged into main`) — the header's `**Branch:**` is\ninformational and the daemon lands only the run branch `swarm\u002Fphase-\u003Cid>`.\nCriteria only a human can perform (production, a console, a manual check)\nstart with `[MANUAL]`. Neither is the executor's to tick or perform: the\ndaemon settles a run against the rest, refuses to start a `[MANUAL]`-only\nphase and lists it in Needs you, lists a phase with only `[LAND]` left as\nawaiting merge, and ticks `[LAND]` itself when the pull request for the run\nbranch merges — whether landed through the Review tab or opened outside the\nplatform and found by the landing poll or \"Find PR\". When the PR comes from\nany other branch, its `[LAND]` criteria are ticked by hand. The header rows\n`**Verify:**` (off | normal | strict), `**Review:**` (on | off) and the required `**Effort:**` (low | medium | high | xhigh)\nsit directly after `**Repo:** …`; L\u002FXL phases default to `normal` \u002F `on`. Details: `resources\u002Fplan-format.md` → \"Criterion classes\"\nand \"Header rows\".\n\nCriteria that lean on the engine follow five rules in `resources\u002Fplan-format.md`:\nspike a guarantee under every permission mode, live-check engine-boundary phases\nbefore `[LAND]`, append live-check evidence immediately, cite\n`resources\u002Fconfig-channels.md` for any settings-file criterion, and number\nmigrations at landing.\n\n# Resources\n\n- `resources\u002Fplan-format.md` — the full field-by-field format, examples, the\n  manifest schema, and the engine-facing criterion rules.\n- `resources\u002Fblast-radius.md` — the planner's blast-radius step: who calls\n  what a phase changes, and the `Blast radius:` line its `## Risks` opens with.\n- `resources\u002Fconfig-channels.md` — which settings tier Claude Code reads for\n  `pluginConfigs`, plugin `env` and `${VAR}`, with the CLI version measured.\n- `templates\u002Fspike-matrix.md` (plugin root) — the permission mode × surface\n  matrix for a spike.\n\n# Executor duties\n\nAfter reading the code the phase touches and before your first edit, add a\n`kind: posterior` block beside the prior in `## Forecast` — the\nprediction-not-a-limit rule above applies unchanged — and if the work then\ndiverges from it, say how and why in a short \"Where reality diverged\"\nparagraph of the Completion Report.\n\nTick each satisfied criterion `- [ ]` → `- [x]` immediately after verifying\nit — progress is derived only from these checkboxes. When a phase's last\nEXECUTABLE criterion is ticked (every one that is not `[LAND]`\u002F`[MANUAL]` —\nthose close later, without an executor), fill that doc's `## Completion Report` (what shipped,\nfiles\u002Fcommits, verification output, deviations, plus three mandatory fields —\n**Blocked calls**, **Delegation cost**, and the one-sentence **What would have\nmade this cheaper**; ≤50 lines) — the dashboard renders exactly that heading.\nEach of the three is written out even when empty (\"none\" \u002F \"Nothing, it was\nalready minimal\"); `resources\u002Fplan-format.md` carries their exact wording and\nis the source of truth. Never tick unsatisfied criteria; never archive\nwith unmet criteria or a missing SUMMARY.md.\n\nFull field-by-field format, examples, and the manifest schema:\n`resources\u002Fplan-format.md`.\n\n# How to use\n\n## What it does\n\nCarries the exact workspace-plan format the dashboard ingests: README sequencing table, phase docs with file-backed-verification executor prompts and checkbox criteria, manifest DAG, optional spec with SC coverage, and the tick\u002FCompletion Report duties for executors — including the report's Blocked calls, Delegation cost and \"what would have made this cheaper\" fields.\n\n## When to use it\n\nLoad it whenever you write a new plan, revise one, or execute one — any time plan files under a task dir's `plan\u002F` are created or updated.\n\n## How to invoke\n\nThe `@core:planner` and `@core:implementation-agent` agents list it in their `skills:`; load it manually with `\u002Fworkspace-plans` context or by reading `resources\u002Fplan-format.md` when editing plan files by hand.\n\n## Worked example\n\nA planner writing `plan\u002Fphase-2-api.md` includes: `**Covers:** SC-3`, a prompt an executor can run cold whose Verify block names `scripts\u002Fcheck-api.sh` rather than inlining a heredoc, three `- [ ]` criteria each with a check command, and ends the doc with an empty `## Completion Report` section. The executor later ticks the boxes one by one and fills the report — Blocked calls, Delegation cost and the one-sentence cheaper field included — and the dashboard's phase Summary tab shows exactly that section.\n",{"docs":24,"name":5,"color":27,"owner":28,"version":29,"description":7},{"status":25,"updated":26},"draft",1792540800000,"cyan","swarmery-core","1.2.0","main","2d9e918cc79c9b17b59356780ff0a3da38a3e98a","plugins\u002Fcore\u002Fskills\u002Fworkspace-plans",5,"https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery","https:\u002F\u002Fatretyak1985.github.io\u002Fswarmery\u002F","https:\u002F\u002Fgithub.com\u002Fatretyak1985","2026-11-11 04:29:12",{"source":8,"skillName":5,"commands":39,"prompt":49,"zipAvailable":50},[40,43,46],{"tool":41,"command":42},"npx","npx skills add atretyak1985\u002Fswarmery --skill workspace-plans",{"tool":44,"command":45},"bunx","bunx skills add atretyak1985\u002Fswarmery --skill workspace-plans",{"tool":47,"command":48},"pnpm","pnpm dlx skills add atretyak1985\u002Fswarmery --skill workspace-plans","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Fplugins\u002Fcore\u002Fskills\u002Fworkspace-plans\u002FSKILL.md",true,[52,58,62],{"path":53,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":56,"sourceHtmlUrl":57},"resources\u002Fblast-radius.md","REFERENCE","参考文档",5177,"https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Fplugins\u002Fcore\u002Fskills\u002Fworkspace-plans\u002Fresources\u002Fblast-radius.md",{"path":59,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":60,"sourceHtmlUrl":61},"resources\u002Fconfig-channels.md",3695,"https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Fplugins\u002Fcore\u002Fskills\u002Fworkspace-plans\u002Fresources\u002Fconfig-channels.md",{"path":63,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":64,"sourceHtmlUrl":65},"resources\u002Fplan-format.md",31609,"https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Fplugins\u002Fcore\u002Fskills\u002Fworkspace-plans\u002Fresources\u002Fplan-format.md",[67,74,80,87,94,101],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M78D4T1FFWFW19Z44AHH5BSY","short-desc","Too short.","https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Ftools\u002Fswarmery\u002Ftestdata\u002Fsysconfig\u002Fclaude\u002Fskills\u002Fshort-desc\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M78D4S0TWT42WXFJHC1DBN0R","no-frontmatter","https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Ftools\u002Fswarmery\u002Ftestdata\u002Fsysconfig\u002Fclaude\u002Fskills\u002Fno-frontmatter\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M78D4Q14P8DJ4W78N62F3EG2","global-skill","Fixture global skill (minimal user-scope shape, format doc §2.3).","https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Ftools\u002Fswarmery\u002Ftestdata\u002Fsysconfig\u002Fclaude\u002Fskills\u002Fglobal-skill\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":19},"01M78D4N1JHGQJQ7EZ43KY5SV3","documented-skill","Fixture skill whose usage guide deliberately omits the Worked example subsection.","https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Ftools\u002Fswarmery\u002Ftestdata\u002Fsysconfig\u002Fclaude\u002Fskills\u002Fdocumented-skill\u002FSKILL.md",[],[],{"skillNo":95,"name":96,"nameZh":6,"description":97,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":98,"categories":99,"keywords":100,"publishedAt":19},"01M78D4KVNZ03GBKGDFJZWW5FJ","z","NOT auto-discovered — only skills\u002F is scanned for plugins (§2.2).","https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Ftools\u002Fswarmery\u002Ftestdata\u002Fsysconfig\u002Fclaude\u002Fplugins\u002Fcache\u002Fmkt\u002Ftoolpack\u002F1.0.0\u002Fworkflow-skills\u002Fz\u002FSKILL.md",[],[],{"skillNo":102,"name":103,"nameZh":6,"description":104,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":105,"categories":106,"keywords":107,"publishedAt":19},"01M78D4JMBRZWK51HFE5RZJGHC","y","Plugin skill y from toolpack — stored as \"toolpack:y\".","https:\u002F\u002Fgithub.com\u002Fatretyak1985\u002Fswarmery\u002Fblob\u002Fmain\u002Ftools\u002Fswarmery\u002Ftestdata\u002Fsysconfig\u002Fclaude\u002Fplugins\u002Fcache\u002Fmkt\u002Ftoolpack\u002F1.0.0\u002Fskills\u002Fy\u002FSKILL.md",[],[]]