[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6DDGVSSK3G29B3THAB0QC6K":3,"skill-same:01M6DDGVSSK3G29B3THAB0QC6K":77},{"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":23,"publishedAt":42,"linkOnly":43,"markdownContent":44,"markdownBody":45,"frontMatter":46,"refName":47,"commitSha":48,"skillDirectory":49,"forksCount":14,"repositoryHtmlUrl":50,"repositoryHomepageUrl":6,"ownerHtmlUrl":51,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":52,"install":53,"resources":66},"01M6DDGVSSK3G29B3THAB0QC6K","mcp-fleet-curation",null,"Use when plugins, skills, or MCP servers keep REAPPEARING after you disable them, when de-bloating the Claude Code catalogue, or when the per-turn tool\u002Ftoken count creeps back up. The permanent cure - disable at USER scope (claude plugin disable name@mkt --scope user), NOT the UI toggle (which writes project scope and returns when the cwd changes) - plus periodic mcp_doctor audit.","Prekzursil\u002Fagent-skills-toolchain","Prekzursil","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F54636077?v=4","01M6DACZ1PCQHD99SCR1YXG28B","Andrei Mihai Visalon","01M6DACZ670WZF535V3WYYSS4Z",0,"MIT","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Fmcp-fleet-curation\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41],"mcp fleet curation","MCP","returns","token","project","back","periodic","up","use","user","tool","ui","audit","plugins","mcp servers","doctor","code","plugin","2026-11-10 10:12:30",false,"---\nname: mcp-fleet-curation\ndescription: Use when plugins, skills, or MCP servers keep REAPPEARING after you disable them, when de-bloating the Claude Code catalogue, or when the per-turn tool\u002Ftoken count creeps back up. The permanent cure - disable at USER scope (claude plugin disable name@mkt --scope user), NOT the UI toggle (which writes project scope and returns when the cwd changes) - plus periodic mcp_doctor audit.\nlicense: MIT\n---\n\n# MCP \u002F Plugin \u002F Skill Fleet Curation\n\n> **Why your disabled plugins\u002FMCPs keep coming back — and the one cure that\n> sticks.** Two bugs, both fixed by disabling at **user scope**. The UI toggle\n> lies to you (it writes *project* scope).\n\n## When to use\n\n- \"I disabled this plugin\u002FMCP and it came back.\"\n- De-bloating the catalogue: too many listed tools\u002Fskills taxing every turn.\n- `\u002Fcontext` shows a creeping always-on token cost.\n- After any `claude plugin install` (curate it immediately — see the rule).\n\n## Root cause (two bugs)\n\n1. **Install = enabled-by-default.** A plugin in `installed_plugins.json` is ON\n   unless an explicit `false` is recorded. Dozens were never explicitly disabled,\n   so every reload re-affirmed them.\n2. **Scope mismatch.** The `\u002Fplugin` UI and a bare `claude plugin disable` write\n   **project** scope → the plugin returns the moment the cwd changes.\n\n## The permanent cure\n\n**Disable at user scope** (writes a durable `false` to `~\u002F.claude\u002Fsettings.json`,\nproven to persist across cwd changes and restarts):\n\n```bash\nclaude plugin disable \u003Cname>@\u003Cmarketplace> --scope user\n```\n\nNever trust the UI toggle for a durable decision. Verify with `\u002Fcontext` after a\nrestart (plugin\u002FMCP changes apply on the next CC start, not live).\n\n## MCP servers: reversible park \u002F restore\n\nMCP servers live in `~\u002F.mcp.json` (active) with a parking lot in\n`~\u002F.mcp.disabled.json`. To **park** a server: move its block out of `mcpServers`\nin `~\u002F.mcp.json` into `~\u002F.mcp.disabled.json`. To **restore**: move the block\nback. Takes effect on the **next** CC restart (servers load at session start;\nalready-open sessions keep their old set). Backups: `~\u002F.mcp.json.bak-*`.\n\nThe project approval-list (`enabledMcpjsonServers` in `settings.local.json`) is a\nsecond gate — trim cross-layer duplicates there too.\n\n## mcp_doctor — catch drift before it snowballs\n\n`~\u002F.local\u002Fopt\u002Fre-harness\u002Fmcp_doctor.py` (+ `lib\u002Fmcp_doctor_core|mcp_inventory|mcp_procscan.py`,\nbudget in `mcp_budget.yaml`):\n\n- `python mcp_doctor.py status` — liveness table for all servers.\n- `python mcp_doctor.py audit` — drift vs the budget manifest (flags re-enabled \u002F\n  duplicate \u002F heavy \u002F per-session pile-up; surfaced docker×25, excel×25, frida×15).\n- `python mcp_doctor.py reap` — registry-gated duplicate-killer (**GUI IDA :13337\n  is protected**; dry-run by default).\n\nRun `audit` periodically — that is the \"curation keeps reappearing\" early-warning.\n\n## The catalogue-tax model (the core confusion)\n\n- **Access = listed in the catalogue** → taxed **every turn**, used or not.\n- **Use = invocation** (`Skill` tool \u002F `ToolSearch`+call \u002F `Agent` tool) → paid\n  only when actually called.\n- `ENABLE_TOOL_SEARCH=auto:5` already **defers** MCP tool schemas (they load on\n  demand). So the remaining lever is **reducing the COUNT** of listed items —\n  exactly what user-scope disabling does. Pair with `mcp-code-orchestrate` to\n  keep *result* tokens down on fan-out, and `ToolSearch` to keep *schema* tokens\n  down on single calls.\n\n## Maintenance rule\n\nEvery new `claude plugin install` is immediately **kept** or\n`disable --scope user`. Never leave it on default-on. Skills under\n`~\u002F.claude\u002Fskills` apply immediately (no restart); plugins\u002FMCPs need a restart.\n\n## Notes \u002F gotchas\n\n- `basic-memory` runs as a shared HTTP server on `:8765` (a single point of\n  failure). If it shows disconnected, run `pwsh -File ~\u002F.basic-memory\u002Fbm-http-server.ps1`\n  to bring it up immediately; a 5-min scheduled task also self-heals it.\n- `D:\\basic-memory\\.hf` (~14 GB) is the GLOBAL HF cache for media\u002FOCR tools, NOT\n  basic-memory's model — do **not** delete it.\n- Authoritative state docs as of writing: `~\u002F.claude\u002FWORKSPACE-CURATION.md` and\n  `~\u002F.claude\u002FMCP-FLEET-IMPROVEMENT-DECISIONS.md`; config backups under\n  `~\u002F.claude\u002F_config-backups\u002F`.\n","# MCP \u002F Plugin \u002F Skill Fleet Curation\n\n> **Why your disabled plugins\u002FMCPs keep coming back — and the one cure that\n> sticks.** Two bugs, both fixed by disabling at **user scope**. The UI toggle\n> lies to you (it writes *project* scope).\n\n## When to use\n\n- \"I disabled this plugin\u002FMCP and it came back.\"\n- De-bloating the catalogue: too many listed tools\u002Fskills taxing every turn.\n- `\u002Fcontext` shows a creeping always-on token cost.\n- After any `claude plugin install` (curate it immediately — see the rule).\n\n## Root cause (two bugs)\n\n1. **Install = enabled-by-default.** A plugin in `installed_plugins.json` is ON\n   unless an explicit `false` is recorded. Dozens were never explicitly disabled,\n   so every reload re-affirmed them.\n2. **Scope mismatch.** The `\u002Fplugin` UI and a bare `claude plugin disable` write\n   **project** scope → the plugin returns the moment the cwd changes.\n\n## The permanent cure\n\n**Disable at user scope** (writes a durable `false` to `~\u002F.claude\u002Fsettings.json`,\nproven to persist across cwd changes and restarts):\n\n```bash\nclaude plugin disable \u003Cname>@\u003Cmarketplace> --scope user\n```\n\nNever trust the UI toggle for a durable decision. Verify with `\u002Fcontext` after a\nrestart (plugin\u002FMCP changes apply on the next CC start, not live).\n\n## MCP servers: reversible park \u002F restore\n\nMCP servers live in `~\u002F.mcp.json` (active) with a parking lot in\n`~\u002F.mcp.disabled.json`. To **park** a server: move its block out of `mcpServers`\nin `~\u002F.mcp.json` into `~\u002F.mcp.disabled.json`. To **restore**: move the block\nback. Takes effect on the **next** CC restart (servers load at session start;\nalready-open sessions keep their old set). Backups: `~\u002F.mcp.json.bak-*`.\n\nThe project approval-list (`enabledMcpjsonServers` in `settings.local.json`) is a\nsecond gate — trim cross-layer duplicates there too.\n\n## mcp_doctor — catch drift before it snowballs\n\n`~\u002F.local\u002Fopt\u002Fre-harness\u002Fmcp_doctor.py` (+ `lib\u002Fmcp_doctor_core|mcp_inventory|mcp_procscan.py`,\nbudget in `mcp_budget.yaml`):\n\n- `python mcp_doctor.py status` — liveness table for all servers.\n- `python mcp_doctor.py audit` — drift vs the budget manifest (flags re-enabled \u002F\n  duplicate \u002F heavy \u002F per-session pile-up; surfaced docker×25, excel×25, frida×15).\n- `python mcp_doctor.py reap` — registry-gated duplicate-killer (**GUI IDA :13337\n  is protected**; dry-run by default).\n\nRun `audit` periodically — that is the \"curation keeps reappearing\" early-warning.\n\n## The catalogue-tax model (the core confusion)\n\n- **Access = listed in the catalogue** → taxed **every turn**, used or not.\n- **Use = invocation** (`Skill` tool \u002F `ToolSearch`+call \u002F `Agent` tool) → paid\n  only when actually called.\n- `ENABLE_TOOL_SEARCH=auto:5` already **defers** MCP tool schemas (they load on\n  demand). So the remaining lever is **reducing the COUNT** of listed items —\n  exactly what user-scope disabling does. Pair with `mcp-code-orchestrate` to\n  keep *result* tokens down on fan-out, and `ToolSearch` to keep *schema* tokens\n  down on single calls.\n\n## Maintenance rule\n\nEvery new `claude plugin install` is immediately **kept** or\n`disable --scope user`. Never leave it on default-on. Skills under\n`~\u002F.claude\u002Fskills` apply immediately (no restart); plugins\u002FMCPs need a restart.\n\n## Notes \u002F gotchas\n\n- `basic-memory` runs as a shared HTTP server on `:8765` (a single point of\n  failure). If it shows disconnected, run `pwsh -File ~\u002F.basic-memory\u002Fbm-http-server.ps1`\n  to bring it up immediately; a 5-min scheduled task also self-heals it.\n- `D:\\basic-memory\\.hf` (~14 GB) is the GLOBAL HF cache for media\u002FOCR tools, NOT\n  basic-memory's model — do **not** delete it.\n- Authoritative state docs as of writing: `~\u002F.claude\u002FWORKSPACE-CURATION.md` and\n  `~\u002F.claude\u002FMCP-FLEET-IMPROVEMENT-DECISIONS.md`; config backups under\n  `~\u002F.claude\u002F_config-backups\u002F`.\n",{"name":5,"license":15,"description":7},"main","a298c929b6cbd662e1fce54302ce520d5ceea1a5","plugin\u002Fskills\u002Fmcp-fleet-curation","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain","https:\u002F\u002Fgithub.com\u002FPrekzursil","2026-11-02 14:30:57",{"source":8,"skillName":5,"commands":54,"prompt":64,"zipAvailable":65},[55,58,61],{"tool":56,"command":57},"npx","npx skills add Prekzursil\u002Fagent-skills-toolchain --skill mcp-fleet-curation",{"tool":59,"command":60},"bunx","bunx skills add Prekzursil\u002Fagent-skills-toolchain --skill mcp-fleet-curation",{"tool":62,"command":63},"pnpm","pnpm dlx skills add Prekzursil\u002Fagent-skills-toolchain --skill mcp-fleet-curation","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Fplugin\u002Fskills\u002Fmcp-fleet-curation\u002FSKILL.md",true,[67,73],{"path":68,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":71,"sourceHtmlUrl":72},"resources\u002FASSESSMENT_CATEGORIES.md","REFERENCE","参考文档",12990,"https:\u002F\u002Fgithub.com\u002Fstrmt7\u002Fai_skills_collection_benchmarked\u002Fblob\u002Fmain\u002Fincluded\u002Fskills\u002Fby-category\u002Fagent-infrastructure-skill-creation\u002Fsecurity-reference\u002Ftoken-integration-analyzer\u002Fresources\u002FASSESSMENT_CATEGORIES.md",{"path":74,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":75,"sourceHtmlUrl":76},"resources\u002FREPORT_TEMPLATES.md",3289,"https:\u002F\u002Fgithub.com\u002Fstrmt7\u002Fai_skills_collection_benchmarked\u002Fblob\u002Fmain\u002Fincluded\u002Fskills\u002Fby-category\u002Fagent-infrastructure-skill-creation\u002Fsecurity-reference\u002Ftoken-integration-analyzer\u002Fresources\u002FREPORT_TEMPLATES.md",[78,99,124,150,163,178],{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":42},"01M6DF7D7478R55P46REYH4QAC","clarify-requirements","Resolve genuine pre-execution ambiguity by scanning repo context, generating 1-6 grounded options, and asking 1-3 focused questions via AskUserQuestion. NOT for mid-task use (injection risk); NOT when the goal is already unambiguous; use to disambiguate vague prompts before any file edits.","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Fworkspace-hooks\u002Fskills\u002Fclarify-requirements\u002FSKILL.md",[],[85,86,87,88,89,90,91,92,93,32,94,95,96,97,98],"clarify requirements","clarify","requirements","risk","options","task","prompts","goal","scanning","context","repo","injection","execution","file",{"skillNo":100,"name":101,"nameZh":6,"description":102,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":103,"categories":104,"keywords":106,"publishedAt":42},"01M6DF79CA4BHJS66J8EZDBPWJ","openmm-md","Run FREE, open-source, Windows-native GPU molecular dynamics via the openmm-md MCP (OpenMM + CUDA on the local NVIDIA GPU). Use for protein \u002F protein-ligand MD when you want a free stack or are on Windows without a Schrodinger license. Covers the full workflow (prepare, solvate, minimize, equilibrate, production, analyze, mmgbsa) AND an honest decision guide vs Schrodinger Desmond, including how to run Desmond legitimately on Linux or a cloud GPU with a real license. This is a deliberately LESSER free alternative and does NOT replace the schrodinger MCP tools. Do NOT use it to pick a free tool when a validated FEP+, WaterMap, IFD-MD, or Prime MM-GBSA result is actually required.","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Fmcp\u002Fopenmm-md-mcp\u002Fskill\u002Fopenmm-md\u002FSKILL.md",[105],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[107,108,109,110,111,112,34,113,114,115,116,117,118,119,32,120,121,122,123,25],"openmm md","how","decision","pick","linux","analyze","stack","open","NVIDIA","gpu","license","workflow","windows","free","molecular dynamics","local","do",{"skillNo":125,"name":126,"nameZh":6,"description":127,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":128,"categories":129,"keywords":136,"publishedAt":42},"01M6DF76C1MG85VEWP8VZCBNMN","visual-review","Take screenshots of web pages and UI using Playwright for visual review and iteration","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Feditor\u002Fmetaswarm\u002Fskills\u002Fvisual-review\u002FSKILL.md",[130,133],{"categoryNo":131,"categoryName":132,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000003","演示文稿",{"categoryNo":134,"categoryName":135,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",[137,138,139,140,141,142,143,144,145,146,108,35,147,148,149],"review","visual","visual review","pages","web","false","iteration","take","check","slides","screenshot","playwright","screenshots",{"skillNo":151,"name":152,"nameZh":6,"description":153,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":154,"categories":155,"keywords":156,"publishedAt":42},"01M6DF73AQVCBZKEN0YXEX9FXP","status","Diagnostic status report — shows metaswarm installation state, project setup, and potential issues","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Feditor\u002Fmetaswarm\u002Fskills\u002Fstatus\u002FSKILL.md",[],[152,157,158,28,159,160,161,162],"setup","report","installation","project setup","issues","state",{"skillNo":164,"name":165,"nameZh":6,"description":166,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":167,"categories":168,"keywords":170,"publishedAt":42},"01M6DF6X6C2XB0P6GX76P7TTYV","start","Use when starting work on any task, when the user mentions metaswarm, or when the user wants to begin tracked development work","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Feditor\u002Fmetaswarm\u002Fskills\u002Fstart\u002FSKILL.md",[169],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[165,171,90,172,173,174,32,33,175,176,177],"start issue","issue","ready","work on issue","work","work on","development",{"skillNo":179,"name":157,"nameZh":6,"description":180,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":181,"categories":182,"keywords":183,"publishedAt":42},"01M6DF6QMJ9AJKFKHHKRT005PQ","Interactive project setup — detects your project, configures metaswarm, writes project-local files","https:\u002F\u002Fgithub.com\u002FPrekzursil\u002Fagent-skills-toolchain\u002Fblob\u002Fmain\u002Ftools\u002Feditor\u002Fmetaswarm\u002Fskills\u002Fsetup\u002FSKILL.md",[],[157,28,122,184,160,185],"files","interactive"]