[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6WYG2RM2DGEVXFANN6RQHB9":3,"skill-same:01M6WYG2RM2DGEVXFANN6RQHB9":86},{"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":66,"commitSha":67,"skillDirectory":68,"forksCount":14,"repositoryHtmlUrl":69,"repositoryHomepageUrl":70,"ownerHtmlUrl":71,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":72,"install":73,"resources":85},"01M6WYG2RM2DGEVXFANN6RQHB9","skill-gate",null,"Session-start meta-skill that enforces the 1% rule: before any action\non a processkit entity or context\u002F file, if there is even a 1% chance\na processkit skill applies, you MUST check skill-finder first. Load\nthis skill at the start of every session that may involve processkit\nwork. Use when beginning a new task, when unsure which skill applies,\nor when you are about to edit context\u002F, create an entity, or run an\nMCP tool.","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\u002Farchive\u002Fprocesskit-install-20261101\u002Fcontext\u002Ftemplates\u002Fprocesskit\u002Fv0.13.0\u002Fcontext\u002Fskills\u002Fprocesskit\u002Fskill-gate\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],"skill gate","gate","domain","when","for","v1","load","which","true","you","new task","MCP","meta-skill","task","use","session","check","tool","Meta","file","2026-11-10 10:12:30",false,"---\nname: skill-gate\ndescription: |\n  Session-start meta-skill that enforces the 1% rule: before any action\n  on a processkit entity or context\u002F file, if there is even a 1% chance\n  a processkit skill applies, you MUST check skill-finder first. Load\n  this skill at the start of every session that may involve processkit\n  work. Use when beginning a new task, when unsure which skill applies,\n  or when you are about to edit context\u002F, create an entity, or run an\n  MCP tool.\nmetadata:\n  processkit:\n    apiVersion: processkit.projectious.work\u002Fv1\n    id: SKILL-skill-gate\n    version: \"1.0.0\"\n    created: 2026-04-11T08:00:00Z\n    category: processkit\n    core: true\n    layer: null\n    uses:\n      - skill: skill-finder\n        purpose: >\n          The mandatory first stop for any processkit domain task —\n          skill-gate enforces that skill-finder is consulted before\n          acting, not after.\n    provides:\n      primitives: []\n      mcp_tools: []\n      assets: []\n---\n\n# Skill Gate\n\n## Intro\n\nskill-gate is the enforcement layer that ensures agents consult the\nprocesskit skill catalog before acting on processkit entities or\n`context\u002F` files. It implements the **1% rule**: if there is even a 1%\nchance a processkit skill covers the current task, you must check\n`skill-finder` before doing anything else. It does not perform work\nitself — it ensures work is performed via the right skill.\n\n## Overview\n\n### The 1% rule\n\nBefore any of the following actions, pause and apply the decision graph:\n\n- Editing or creating any file under `context\u002F`\n- Calling any processkit MCP tool (`create_*`, `transition_*`, `link_*`,\n  `record_*`, `open_*`)\n- Writing a WorkItem, Decision, Note, Artifact, or any other entity\n- Running a process the user described as a processkit task\n\n**The rule:** If there is even a **1% chance** a processkit skill\napplies to this task, you MUST check `skill-finder` before proceeding.\n\nThe threshold is deliberately near-zero. Do not rationalize past it.\n\n### Decision graph\n\n```\nIs this task related to processkit entities, context\u002F files,\nor any process the user described as a project-management task?\n│\n├─ YES (or unsure) ──→ Read skill-finder and identify the matching skill.\n│                       Load that skill. Work from it.\n│\n└─ NO ──────────────→ Proceed without skill-gate involvement.\n                       (Pure code edits, data analysis, etc. that have\n                       no processkit component do not need skill-gate.)\n```\n\n### What \"check skill-finder\" means\n\n1. Read `context\u002Fskills\u002Fprocesskit\u002Fskill-finder\u002FSKILL.md` — find the\n   trigger phrase that matches the current task.\n2. Load the identified skill's `SKILL.md`.\n3. Confirm from its **Intro** (3 sentences) that it is the right match.\n4. Work from that skill's **Overview** — not from general knowledge.\n\nIf `skill-finder` has an MCP server available (`find_skill` tool), call\nit instead of reading the file — it is faster and more reliable.\n\n### Pre-empting common rationalisations\n\nThese are the rationalizations agents use to skip skill-gate. Each one\nis wrong:\n\n| Rationalisation | Reality |\n|---|---|\n| \"I already know how to do this.\" | The skill may have processkit-specific entity paths, ID formats, MCP tools, and log conventions that general knowledge does not cover. Knowing the domain is not the same as knowing the processkit conventions for it. |\n| \"It's a simple edit — one line.\" | The state machine enforces valid transitions. A hand edit that bypasses the MCP server can put an entity in an invalid state, break the index, and skip the auto-log side effect. |\n| \"skill-finder won't have a match for this.\" | The catalog has 128+ skills. Check before assuming. If there genuinely is no match, `skill-builder` is the next step — not acting from general knowledge. |\n| \"The user told me exactly what to do.\" | User instructions describe intent, not process. The skill encodes the processkit-specific conventions for realizing that intent correctly. |\n| \"I've done this before in this session.\" | Skill-gate applies to each new task type, not just the first one. If the task domain shifts, check again. |\n\n### Escape hatch\n\nskill-gate does NOT apply when you are already operating inside a\nnamed processkit skill's workflow. If you loaded `workitem-management`\nand are following its Overview steps, you do not need to re-check\nskill-finder for each sub-action within that skill. Skill-gate is a\nsession-level gate, not a per-tool-call gate.\n\n## Gotchas\n\n- **Treating skill-gate as optional at session start.** It is not. If\n  this session may involve any processkit work, load skill-gate first.\n  The cost is one file read. The cost of skipping it is a hand edit\n  that bypasses schema validation, state-machine enforcement, index\n  sync, and the audit trail.\n- **Applying the 1% rule only to obvious processkit tasks.** \"Write a\n  standup\" and \"create a ticket\" are obvious. \"Quick note about this\n  idea\" (→ `note-management`), \"log that we deployed\" (→ `event-log`),\n  and \"check what's blocking us\" (→ `workitem-management`) are not\n  obvious but all have matching skills. When in doubt, check.\n- **Confusing skill-gate with skill-finder.** skill-gate enforces the\n  checking behaviour. skill-finder does the matching. They are distinct.\n  skill-gate tells you *when* to check; skill-finder tells you *what*\n  to use.\n- **Skipping skill-gate because the user said \"just edit the file.\"**\n  The user describes the outcome they want, not the correct processkit\n  process for achieving it. Hand-editing entity files bypasses the MCP\n  server's schema validation and state-machine enforcement. Use the MCP\n  tool unless the MCP server is explicitly unavailable.\n- **Forgetting that skill-gate applies to entity creation, not just\n  edits.** Writing a new YAML file under `context\u002Fworkitems\u002F` directly\n  is a hand edit. Even for new entities, use `create_workitem` (or the\n  relevant `create_*` tool) so the index stays in sync and the auto-log\n  fires.\n- **Applying the escape hatch too broadly.** The escape hatch covers\n  sub-actions *within* a skill you have already loaded and are actively\n  following. It does not cover switching to a different task type or\n  domain mid-session without re-checking.\n\n## Full reference\n\n### Why the threshold is 1%, not \"when clearly applicable\"\n\nA higher threshold — \"check skill-finder when the task is obviously a\nprocesskit task\" — fails because agents are systematically overconfident\nabout what they already know. The 1% rule removes the judgment call\nentirely. If there is any possibility, check. The marginal cost of an\nunnecessary skill-finder lookup is one file read (seconds). The cost of\nskipping it when it was needed is an out-of-sync index, an invalid\nentity state, or a missing audit trail entry.\n\n### Provider neutrality\n\nskill-gate is implemented entirely in prose — it makes no assumptions\nabout which AI harness loads it. The same behaviour applies whether the\nagent is Claude Code, Codex CLI, Cursor, or any other harness that\nreads SKILL.md files.\n\n### Relationship to AGENTS.md\n\nAGENTS.md carries the project-level instruction to load skill-gate\nat session start. skill-gate carries the enforcement logic. Keep them\nseparate: AGENTS.md says *when* to load skill-gate; skill-gate says\n*what to do* once loaded.\n\n### When skill-gate applies to non-processkit projects\n\nskill-gate is a processkit-specific skill. It is only meaningful in\nprojects that have processkit installed (i.e. that have a `context\u002F`\ndirectory with processkit entities). In projects without processkit,\nskill-gate should not be loaded.\n","# Skill Gate\n\n## Intro\n\nskill-gate is the enforcement layer that ensures agents consult the\nprocesskit skill catalog before acting on processkit entities or\n`context\u002F` files. It implements the **1% rule**: if there is even a 1%\nchance a processkit skill covers the current task, you must check\n`skill-finder` before doing anything else. It does not perform work\nitself — it ensures work is performed via the right skill.\n\n## Overview\n\n### The 1% rule\n\nBefore any of the following actions, pause and apply the decision graph:\n\n- Editing or creating any file under `context\u002F`\n- Calling any processkit MCP tool (`create_*`, `transition_*`, `link_*`,\n  `record_*`, `open_*`)\n- Writing a WorkItem, Decision, Note, Artifact, or any other entity\n- Running a process the user described as a processkit task\n\n**The rule:** If there is even a **1% chance** a processkit skill\napplies to this task, you MUST check `skill-finder` before proceeding.\n\nThe threshold is deliberately near-zero. Do not rationalize past it.\n\n### Decision graph\n\n```\nIs this task related to processkit entities, context\u002F files,\nor any process the user described as a project-management task?\n│\n├─ YES (or unsure) ──→ Read skill-finder and identify the matching skill.\n│                       Load that skill. Work from it.\n│\n└─ NO ──────────────→ Proceed without skill-gate involvement.\n                       (Pure code edits, data analysis, etc. that have\n                       no processkit component do not need skill-gate.)\n```\n\n### What \"check skill-finder\" means\n\n1. Read `context\u002Fskills\u002Fprocesskit\u002Fskill-finder\u002FSKILL.md` — find the\n   trigger phrase that matches the current task.\n2. Load the identified skill's `SKILL.md`.\n3. Confirm from its **Intro** (3 sentences) that it is the right match.\n4. Work from that skill's **Overview** — not from general knowledge.\n\nIf `skill-finder` has an MCP server available (`find_skill` tool), call\nit instead of reading the file — it is faster and more reliable.\n\n### Pre-empting common rationalisations\n\nThese are the rationalizations agents use to skip skill-gate. Each one\nis wrong:\n\n| Rationalisation | Reality |\n|---|---|\n| \"I already know how to do this.\" | The skill may have processkit-specific entity paths, ID formats, MCP tools, and log conventions that general knowledge does not cover. Knowing the domain is not the same as knowing the processkit conventions for it. |\n| \"It's a simple edit — one line.\" | The state machine enforces valid transitions. A hand edit that bypasses the MCP server can put an entity in an invalid state, break the index, and skip the auto-log side effect. |\n| \"skill-finder won't have a match for this.\" | The catalog has 128+ skills. Check before assuming. If there genuinely is no match, `skill-builder` is the next step — not acting from general knowledge. |\n| \"The user told me exactly what to do.\" | User instructions describe intent, not process. The skill encodes the processkit-specific conventions for realizing that intent correctly. |\n| \"I've done this before in this session.\" | Skill-gate applies to each new task type, not just the first one. If the task domain shifts, check again. |\n\n### Escape hatch\n\nskill-gate does NOT apply when you are already operating inside a\nnamed processkit skill's workflow. If you loaded `workitem-management`\nand are following its Overview steps, you do not need to re-check\nskill-finder for each sub-action within that skill. Skill-gate is a\nsession-level gate, not a per-tool-call gate.\n\n## Gotchas\n\n- **Treating skill-gate as optional at session start.** It is not. If\n  this session may involve any processkit work, load skill-gate first.\n  The cost is one file read. The cost of skipping it is a hand edit\n  that bypasses schema validation, state-machine enforcement, index\n  sync, and the audit trail.\n- **Applying the 1% rule only to obvious processkit tasks.** \"Write a\n  standup\" and \"create a ticket\" are obvious. \"Quick note about this\n  idea\" (→ `note-management`), \"log that we deployed\" (→ `event-log`),\n  and \"check what's blocking us\" (→ `workitem-management`) are not\n  obvious but all have matching skills. When in doubt, check.\n- **Confusing skill-gate with skill-finder.** skill-gate enforces the\n  checking behaviour. skill-finder does the matching. They are distinct.\n  skill-gate tells you *when* to check; skill-finder tells you *what*\n  to use.\n- **Skipping skill-gate because the user said \"just edit the file.\"**\n  The user describes the outcome they want, not the correct processkit\n  process for achieving it. Hand-editing entity files bypasses the MCP\n  server's schema validation and state-machine enforcement. Use the MCP\n  tool unless the MCP server is explicitly unavailable.\n- **Forgetting that skill-gate applies to entity creation, not just\n  edits.** Writing a new YAML file under `context\u002Fworkitems\u002F` directly\n  is a hand edit. Even for new entities, use `create_workitem` (or the\n  relevant `create_*` tool) so the index stays in sync and the auto-log\n  fires.\n- **Applying the escape hatch too broadly.** The escape hatch covers\n  sub-actions *within* a skill you have already loaded and are actively\n  following. It does not cover switching to a different task type or\n  domain mid-session without re-checking.\n\n## Full reference\n\n### Why the threshold is 1%, not \"when clearly applicable\"\n\nA higher threshold — \"check skill-finder when the task is obviously a\nprocesskit task\" — fails because agents are systematically overconfident\nabout what they already know. The 1% rule removes the judgment call\nentirely. If there is any possibility, check. The marginal cost of an\nunnecessary skill-finder lookup is one file read (seconds). The cost of\nskipping it when it was needed is an out-of-sync index, an invalid\nentity state, or a missing audit trail entry.\n\n### Provider neutrality\n\nskill-gate is implemented entirely in prose — it makes no assumptions\nabout which AI harness loads it. The same behaviour applies whether the\nagent is Claude Code, Codex CLI, Cursor, or any other harness that\nreads SKILL.md files.\n\n### Relationship to AGENTS.md\n\nAGENTS.md carries the project-level instruction to load skill-gate\nat session start. skill-gate carries the enforcement logic. Keep them\nseparate: AGENTS.md says *when* to load skill-gate; skill-gate says\n*what to do* once loaded.\n\n### When skill-gate applies to non-processkit projects\n\nskill-gate is a processkit-specific skill. It is only meaningful in\nprojects that have processkit installed (i.e. that have a `context\u002F`\ndirectory with processkit entities). In projects without processkit,\nskill-gate should not be loaded.\n",{"name":5,"metadata":49,"description":65},{"processkit":50},{"id":51,"core":52,"uses":53,"layer":6,"created":57,"version":58,"category":59,"provides":60,"apiVersion":64},"SKILL-skill-gate",true,[54],{"skill":55,"purpose":56},"skill-finder","The mandatory first stop for any processkit domain task — skill-gate enforces that skill-finder is consulted before acting, not after.\n",1775894400000,"1.0.0","processkit",{"assets":61,"mcp_tools":62,"primitives":63},[],[],[],"processkit.projectious.work\u002Fv1","Session-start meta-skill that enforces the 1% rule: before any action\non a processkit entity or context\u002F file, if there is even a 1% chance\na processkit skill applies, you MUST check skill-finder first. Load\nthis skill at the start of every session that may involve processkit\nwork. Use when beginning a new task, when unsure which skill applies,\nor when you are about to edit context\u002F, create an entity, or run an\nMCP tool.\n","main","4876a92620ee792aac2ac9a1a806322d8a572367","archive\u002Fprocesskit-install-20261101\u002Fcontext\u002Ftemplates\u002Fprocesskit\u002Fv0.13.0\u002Fcontext\u002Fskills\u002Fprocesskit\u002Fskill-gate","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":74,"prompt":84,"zipAvailable":52},[75,78,81],{"tool":76,"command":77},"npx","npx skills add projectious-work\u002Fprocesskit --skill skill-gate",{"tool":79,"command":80},"bunx","bunx skills add projectious-work\u002Fprocesskit --skill skill-gate",{"tool":82,"command":83},"pnpm","pnpm dlx skills add projectious-work\u002Fprocesskit --skill skill-gate","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002Farchive\u002Fprocesskit-install-20261101\u002Fcontext\u002Ftemplates\u002Fprocesskit\u002Fv0.13.0\u002Fcontext\u002Fskills\u002Fprocesskit\u002Fskill-gate\u002FSKILL.md",[],[87,94,101,107,130,149],{"skillNo":88,"name":89,"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":90,"categories":91,"keywords":92,"publishedAt":93},"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":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":93},"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":102,"name":103,"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":104,"categories":105,"keywords":106,"publishedAt":93},"01M76650NN4T8P241021NBR64S","standard-agent-building","https:\u002F\u002Fgithub.com\u002Fprojectious-work\u002Fprocesskit\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fstandard-agent-building\u002FSKILL.md",[],[],{"skillNo":108,"name":109,"nameZh":6,"description":110,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":111,"categories":112,"keywords":114,"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",[113],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[115,116,117,29,118,119,120,121,122,123,124,125,109,126,127,128,129],"management","skill management","ontology","create","as","work","transition","get","manage","update","alpha","list","package","skill package","manifests",{"skillNo":131,"name":132,"nameZh":6,"description":133,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":134,"categories":135,"keywords":137,"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",[136],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[138,139,115,140,141,142,143,144,145,35,146,147,120,122,123,125,148],"schema","schema management","contract","rebuild","contracts","workflows","wed","generated","validation","deterministic","mode",{"skillNo":150,"name":151,"nameZh":6,"description":152,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":153,"categories":154,"keywords":155,"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",[],[156,115,157,158,118,159,160,161,120,122,162,163,124,164,125,165,166,167,168,169],"proposition management","proposition","interface","claims","record","world","creation","event","WSJF","index","updates","query","facts","log"]