[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6NQ8MAY530C34GXS048TFCB":3,"skill-same:01M6NQ8MAY530C34GXS048TFCB":103},{"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":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":67,"commitSha":68,"skillDirectory":69,"forksCount":14,"repositoryHtmlUrl":70,"repositoryHomepageUrl":71,"ownerHtmlUrl":72,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":73,"install":74,"resources":87},"01M6NQ8MAY530C34GXS048TFCB","aggregate-mcp",null,"Expose the processkit MCP tool surface through a single stdio server.\nUse for harnesses that eagerly start every configured MCP server, such\nas Codex, while keeping per-skill MCP servers available.","projectious-work\u002Fprocesskit","projectious-work","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F263889726?v=4","01M6N97TJJ118WQ35CHRV8RD0E","projectious.work","01M6N97TPWJPF0BMKYQ4PBY5W8",0,"MIT","https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002Fsrc\u002Fcontext\u002Fskills\u002Fprocesskit\u002Faggregate-mcp\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"aggregate mcp","MCP","work","servers","for","harnesses","as","imported","wed","use","list","json","tool","start","server","index","mcp servers","config","mcp server","codex","2026-11-10 10:12:30",false,"---\nname: aggregate-mcp\ndescription: |\n  Expose the processkit MCP tool surface through a single stdio server.\n  Use for harnesses that eagerly start every configured MCP server, such\n  as Codex, while keeping per-skill MCP servers available.\nmetadata:\n  processkit:\n    apiVersion: processkit.projectious.work\u002Fv2\n    id: SKILL-aggregate-mcp\n    version: \"1.0.0\"\n    created: 2026-04-29T00:00:00Z\n    category: processkit\n    layer: 1\n    uses:\n      - skill: index-management\n        purpose: Imported as one of the aggregate MCP tool providers.\n    provides:\n      mcp_tools: [list_aggregate_tools]\n      opt_in_config:\n        - mcp\u002Fmcp-config.aggregate.json\n---\n\n# Aggregate MCP\n\n## Intro\n\nThis skill ships a single MCP server that imports the processkit per-skill\nMCP servers and re-exports their tools through one stdio process. It is a\nstartup-latency optimization for harnesses that eagerly initialize every\nconfigured MCP server.\n\nPer-skill servers remain canonical and should continue to be shipped. The\naggregate server is an alternate entry point for clients that prefer one\nprocess over many.\n\n## Overview\n\nUse this server when a harness pays a large startup cost for many stdio\nMCP processes. The aggregate server imports the same tool functions that\nthe granular processkit servers expose, so the processkit write\u002Fread\nsemantics stay in one code path.\n\nThe runtime supports two import modes:\n\n- `eager` (default): importing `aggregate-mcp\u002Fmcp\u002Fserver.py` imports\n  every processkit per-skill MCP server and registers the discovered\n  tools immediately. Backward-compatible with earlier releases.\n- `lazy_catalog`: opt in with `PROCESSKIT_MCP_LAZY=1` (or\n  `PROCESSKIT_MCP_MODE=lazy_catalog`). The aggregate server loads the\n  shared `processkit-gateway\u002Fmcp\u002Ftool-catalog.json` manifest at\n  startup and announces the full tool surface from cached metadata.\n  The owning per-skill module is imported only when one of its tools\n  is first called. Override the catalog source with\n  `PROCESSKIT_MCP_CATALOG_PATH=\u002Fabs\u002Fpath\u002Fto\u002Ftool-catalog.json` if\n  needed. If the catalog is missing the server falls back to eager\n  mode automatically.\n\nCache invalidation is build-time: regenerate the catalog with\n`processkit-gateway catalog --write` (gateway server CLI) and commit\nthe refreshed `tool-catalog.json`. Lazy mode does not regenerate at\nstartup — that would defeat the cold-start win.\n\n## Tool Names\n\nTools keep their original names when the name is unique across processkit.\nIf two source servers expose the same tool name, the aggregate server keeps\nthe first one and registers later duplicates as\n`\u003Cskill_slug>__\u003Ctool_name>`. This preserves familiar names such as\n`create_workitem` while keeping duplicate helpers like `reload_schemas`\navailable.\n\n`list_aggregate_tools` is the registry surface for gateway adapters. It\nreports each registered tool's aggregate name, source skill, source MCP\nserver path, source tool name, serialized annotations, collision status,\ncollision source list, and whether the aggregate name was deduplicated.\nThe response also includes runtime metadata:\n\n- `runtime.import_mode`: one of `eager`, `gateway_registry`,\n  `lazy_catalog`\n- `runtime.lazy_daemon`: `true` only in `lazy_catalog` mode\n- `runtime.registry_backend`: e.g. `eager_import`,\n  `gateway_registry`, `gateway_lazy_catalog`\n- `runtime.canonical_gateway_available`: `true` when the shared\n  `processkit.gateway` library is importable\n- `runtime.catalog_path`: present only in `lazy_catalog` mode\n- `source_server_count`\n\nCollision status values are:\n\n- `unique` - no other source server exports that tool name.\n- `primary` - the first source server kept the unprefixed tool name.\n- `namespaced_duplicate` - a later source server was registered with a\n  source-skill prefix.\n\n## Aggregate Mode Config\n\nThe opt-in fragment at `mcp\u002Fmcp-config.aggregate.json` registers only\n`processkit-aggregate-mcp`:\n\n```json\n{\n  \"mcpServers\": {\n    \"processkit-aggregate-mcp\": {\n      \"command\": \"uv\",\n      \"args\": [\n        \"run\",\n        \"context\u002Fskills\u002Fprocesskit\u002Faggregate-mcp\u002Fmcp\u002Fserver.py\"\n      ],\n      \"env\": {\n        \"PROCESSKIT_MCP_MODE\": \"aggregate\"\n      }\n    }\n  }\n}\n```\n\nHarness adapters such as aibox should merge this fragment instead of\nthe per-skill `mcp-config.json` fragments when aggregate mode is\nenabled. It is intentionally not named `mcp-config.json`, so existing\ngranular installers do not register aggregate mode beside all granular\nservers and increase startup work.\n\n## Gotchas\n\n- Do not remove the per-skill MCP servers. The aggregate server is an\n  adapter for eager-start clients, not the canonical implementation.\n- Harness adapters must opt into this server deliberately. Projects that\n  rely on fine-grained MCP server permissions may still prefer per-skill\n  registration.\n- Duplicate helper tools are namespaced only when needed. Callers should\n  prefer the original unprefixed names for unique tools.\n- `lazy_catalog` mode requires the cached catalog under\n  `processkit-gateway\u002Fmcp\u002Ftool-catalog.json`. If a skill's tools are\n  not in the catalog they will not be exposed by the lazy server until\n  the catalog is regenerated. Run `processkit-gateway catalog --write`\n  after adding or renaming any per-skill tool.\n- The full daemon gateway (long-lived process, streamable-http\n  transport) lives in the `processkit-gateway` skill. The aggregate\n  server is the lighter-weight stdio surface and now shares the same\n  catalog and registration internals.\n\n## Full reference\n\nThe aggregate server lives at `mcp\u002Fserver.py`. It intentionally does\nnot ship a default `mcp-config.json`, because registering it beside the\ngranular servers would increase startup work instead of reducing it.\nHarness adapters can use `mcp\u002Fmcp-config.aggregate.json` as the opt-in\nreplacement fragment when they support aggregate mode:\n\n```json\n{\n  \"mcpServers\": {\n    \"processkit-aggregate-mcp\": {\n      \"command\": \"uv\",\n      \"args\": [\n        \"run\",\n        \"context\u002Fskills\u002Fprocesskit\u002Faggregate-mcp\u002Fmcp\u002Fserver.py\"\n      ],\n      \"env\": {\n        \"PROCESSKIT_MCP_MODE\": \"aggregate\"\n      }\n    }\n  }\n}\n```\n\nThe `list_aggregate_tools` tool reports the source skill, source server\npath, source tool, annotations, and collision status for every exported\ntool.\n","# Aggregate MCP\n\n## Intro\n\nThis skill ships a single MCP server that imports the processkit per-skill\nMCP servers and re-exports their tools through one stdio process. It is a\nstartup-latency optimization for harnesses that eagerly initialize every\nconfigured MCP server.\n\nPer-skill servers remain canonical and should continue to be shipped. The\naggregate server is an alternate entry point for clients that prefer one\nprocess over many.\n\n## Overview\n\nUse this server when a harness pays a large startup cost for many stdio\nMCP processes. The aggregate server imports the same tool functions that\nthe granular processkit servers expose, so the processkit write\u002Fread\nsemantics stay in one code path.\n\nThe runtime supports two import modes:\n\n- `eager` (default): importing `aggregate-mcp\u002Fmcp\u002Fserver.py` imports\n  every processkit per-skill MCP server and registers the discovered\n  tools immediately. Backward-compatible with earlier releases.\n- `lazy_catalog`: opt in with `PROCESSKIT_MCP_LAZY=1` (or\n  `PROCESSKIT_MCP_MODE=lazy_catalog`). The aggregate server loads the\n  shared `processkit-gateway\u002Fmcp\u002Ftool-catalog.json` manifest at\n  startup and announces the full tool surface from cached metadata.\n  The owning per-skill module is imported only when one of its tools\n  is first called. Override the catalog source with\n  `PROCESSKIT_MCP_CATALOG_PATH=\u002Fabs\u002Fpath\u002Fto\u002Ftool-catalog.json` if\n  needed. If the catalog is missing the server falls back to eager\n  mode automatically.\n\nCache invalidation is build-time: regenerate the catalog with\n`processkit-gateway catalog --write` (gateway server CLI) and commit\nthe refreshed `tool-catalog.json`. Lazy mode does not regenerate at\nstartup — that would defeat the cold-start win.\n\n## Tool Names\n\nTools keep their original names when the name is unique across processkit.\nIf two source servers expose the same tool name, the aggregate server keeps\nthe first one and registers later duplicates as\n`\u003Cskill_slug>__\u003Ctool_name>`. This preserves familiar names such as\n`create_workitem` while keeping duplicate helpers like `reload_schemas`\navailable.\n\n`list_aggregate_tools` is the registry surface for gateway adapters. It\nreports each registered tool's aggregate name, source skill, source MCP\nserver path, source tool name, serialized annotations, collision status,\ncollision source list, and whether the aggregate name was deduplicated.\nThe response also includes runtime metadata:\n\n- `runtime.import_mode`: one of `eager`, `gateway_registry`,\n  `lazy_catalog`\n- `runtime.lazy_daemon`: `true` only in `lazy_catalog` mode\n- `runtime.registry_backend`: e.g. `eager_import`,\n  `gateway_registry`, `gateway_lazy_catalog`\n- `runtime.canonical_gateway_available`: `true` when the shared\n  `processkit.gateway` library is importable\n- `runtime.catalog_path`: present only in `lazy_catalog` mode\n- `source_server_count`\n\nCollision status values are:\n\n- `unique` - no other source server exports that tool name.\n- `primary` - the first source server kept the unprefixed tool name.\n- `namespaced_duplicate` - a later source server was registered with a\n  source-skill prefix.\n\n## Aggregate Mode Config\n\nThe opt-in fragment at `mcp\u002Fmcp-config.aggregate.json` registers only\n`processkit-aggregate-mcp`:\n\n```json\n{\n  \"mcpServers\": {\n    \"processkit-aggregate-mcp\": {\n      \"command\": \"uv\",\n      \"args\": [\n        \"run\",\n        \"context\u002Fskills\u002Fprocesskit\u002Faggregate-mcp\u002Fmcp\u002Fserver.py\"\n      ],\n      \"env\": {\n        \"PROCESSKIT_MCP_MODE\": \"aggregate\"\n      }\n    }\n  }\n}\n```\n\nHarness adapters such as aibox should merge this fragment instead of\nthe per-skill `mcp-config.json` fragments when aggregate mode is\nenabled. It is intentionally not named `mcp-config.json`, so existing\ngranular installers do not register aggregate mode beside all granular\nservers and increase startup work.\n\n## Gotchas\n\n- Do not remove the per-skill MCP servers. The aggregate server is an\n  adapter for eager-start clients, not the canonical implementation.\n- Harness adapters must opt into this server deliberately. Projects that\n  rely on fine-grained MCP server permissions may still prefer per-skill\n  registration.\n- Duplicate helper tools are namespaced only when needed. Callers should\n  prefer the original unprefixed names for unique tools.\n- `lazy_catalog` mode requires the cached catalog under\n  `processkit-gateway\u002Fmcp\u002Ftool-catalog.json`. If a skill's tools are\n  not in the catalog they will not be exposed by the lazy server until\n  the catalog is regenerated. Run `processkit-gateway catalog --write`\n  after adding or renaming any per-skill tool.\n- The full daemon gateway (long-lived process, streamable-http\n  transport) lives in the `processkit-gateway` skill. The aggregate\n  server is the lighter-weight stdio surface and now shares the same\n  catalog and registration internals.\n\n## Full reference\n\nThe aggregate server lives at `mcp\u002Fserver.py`. It intentionally does\nnot ship a default `mcp-config.json`, because registering it beside the\ngranular servers would increase startup work instead of reducing it.\nHarness adapters can use `mcp\u002Fmcp-config.aggregate.json` as the opt-in\nreplacement fragment when they support aggregate mode:\n\n```json\n{\n  \"mcpServers\": {\n    \"processkit-aggregate-mcp\": {\n      \"command\": \"uv\",\n      \"args\": [\n        \"run\",\n        \"context\u002Fskills\u002Fprocesskit\u002Faggregate-mcp\u002Fmcp\u002Fserver.py\"\n      ],\n      \"env\": {\n        \"PROCESSKIT_MCP_MODE\": \"aggregate\"\n      }\n    }\n  }\n}\n```\n\nThe `list_aggregate_tools` tool reports the source skill, source server\npath, source tool, annotations, and collision status for every exported\ntool.\n",{"name":5,"metadata":49,"description":66},{"processkit":50},{"id":51,"uses":52,"layer":56,"created":57,"version":58,"category":59,"provides":60,"apiVersion":65},"SKILL-aggregate-mcp",[53],{"skill":54,"purpose":55},"index-management","Imported as one of the aggregate MCP tool providers.",1,1777420800000,"1.0.0","processkit",{"mcp_tools":61,"opt_in_config":63},[62],"list_aggregate_tools",[64],"mcp\u002Fmcp-config.aggregate.json","processkit.projectious.work\u002Fv2","Expose the processkit MCP tool surface through a single stdio server.\nUse for harnesses that eagerly start every configured MCP server, such\nas Codex, while keeping per-skill MCP servers available.\n","main","7231f100e266e30d2c1699ecde5c00602471dc3c","src\u002Fcontext\u002Fskills\u002Fprocesskit\u002Faggregate-mcp","https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit","https:\u002F\u002Fprojectious-work.github.io\u002Fprocesskit\u002F","https:\u002F\u002Fgithub.com\u002Fprojectious-work","2026-11-11 23:24:18",{"source":8,"skillName":5,"commands":75,"prompt":85,"zipAvailable":86},[76,79,82],{"tool":77,"command":78},"npx","npx skills add projectious-work\u002Fprocesskit --skill aggregate-mcp",{"tool":80,"command":81},"bunx","bunx skills add projectious-work\u002Fprocesskit --skill aggregate-mcp",{"tool":83,"command":84},"pnpm","pnpm dlx skills add projectious-work\u002Fprocesskit --skill aggregate-mcp","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002Fsrc\u002Fcontext\u002Fskills\u002Fprocesskit\u002Faggregate-mcp\u002FSKILL.md",true,[88,93,99],{"path":64,"resourceType":89,"resourceTypeDesc":90,"sizeBytes":91,"sourceHtmlUrl":92},"REFERENCE","参考文档",259,"https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002Fsrc\u002Fcontext\u002Fskills\u002Fprocesskit\u002Faggregate-mcp\u002Fmcp\u002Fmcp-config.aggregate.json",{"path":94,"resourceType":95,"resourceTypeDesc":96,"sizeBytes":97,"sourceHtmlUrl":98},"mcp\u002Fserver.py","SCRIPT","脚本",15439,"https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002Fsrc\u002Fcontext\u002Fskills\u002Fprocesskit\u002Faggregate-mcp\u002Fmcp\u002Fserver.py",{"path":100,"resourceType":95,"resourceTypeDesc":96,"sizeBytes":101,"sourceHtmlUrl":102},"scripts\u002Ftest_aggregate_mcp.py",9723,"https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002Fsrc\u002Fcontext\u002Fskills\u002Fprocesskit\u002Faggregate-mcp\u002Fscripts\u002Ftest_aggregate_mcp.py",[104,111,118,124,145,163],{"skillNo":105,"name":106,"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":107,"categories":108,"keywords":109,"publishedAt":110},"01M7665G2ED3624R3X93304265","standard-ui-design","https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fstandard-ui-design\u002FSKILL.md",[],[],"2026-11-12 13:55:43",{"skillNo":112,"name":113,"nameZh":6,"description":114,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":115,"categories":116,"keywords":117,"publishedAt":110},"01M76651QTC6W0HJ4AFW2MMR3V","standard-agent-use","Company standard for working with AI agents. Use when deciding what to delegate to agents, capturing open work, keeping agent memory or knowledge notes, improving an agent workflow, or introducing agent work in a team or for a client.","https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fstandard-agent-use\u002FSKILL.md",[],[],{"skillNo":119,"name":120,"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":121,"categories":122,"keywords":123,"publishedAt":110},"01M76650NN4T8P241021NBR64S","standard-agent-building","https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fstandard-agent-building\u002FSKILL.md",[],[],{"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":131,"publishedAt":44},"01M6X5MTC89B4ZGXW3JVSQCPF8","skill-management","Manage Skill package manifests as v1 ontology entities.","https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002Fsrc\u002Fcontext\u002Fskills\u002Fprocesskit\u002Fskill-management\u002FSKILL.md",[130],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[132,133,134,135,136,30,26,137,138,139,140,141,126,34,142,143,144],"management","skill management","ontology","v1","create","transition","get","manage","update","alpha","package","skill package","manifests",{"skillNo":146,"name":147,"nameZh":6,"description":148,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":149,"categories":150,"keywords":152,"publishedAt":44},"01M6X5MCV74NHW5MEDNHZRF2C8","schema-management","Manage generated processkit schema contracts through deterministic MCP rebuild and inspection workflows.","https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002Fsrc\u002Fcontext\u002Fskills\u002Fprocesskit\u002Fschema-management\u002FSKILL.md",[151],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[153,154,132,155,156,157,158,32,159,25,160,161,26,138,139,141,162],"schema","schema management","contract","rebuild","contracts","workflows","generated","validation","deterministic","mode",{"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":169,"publishedAt":44},"01M6X5M5G6YJ96AMZGSCZJ1D13","proposition-management","Record claims, risks, beliefs, world facts, WSJF estimates, and assumptions through the Proposition interface.","https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002Fsrc\u002Fcontext\u002Fskills\u002Fprocesskit\u002Fproposition-management\u002FSKILL.md",[],[170,132,171,172,136,173,174,175,26,138,176,177,140,178,141,39,179,180,181,182],"proposition management","proposition","interface","claims","record","world","creation","event","WSJF","updates","query","facts","log"]