[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77K381EHV79HSCT4VETKAMG":3,"skill-same:01M77K381EHV79HSCT4VETKAMG":59},{"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":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":6,"ownerHtmlUrl":32,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":33,"install":34,"resources":46},"01M77K381EHV79HSCT4VETKAMG","skill-create",null,"Scaffold a new skill in the monorepo with a valid manifest and catalog entry. Use when creating any new skill.","Flow-Research\u002Fharnessy-v2","Flow-Research","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F204254796?v=4","01M6GVJZH9ZV0XGZTVBJRNW5BQ","Flow Research","01M77JYBHV7WQEBVS6R0AQ4YT6",13,"MIT","https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fcapability-harnessy-v1-full\u002Fresources\u002Fsource\u002Ftools\u002Fflow-install\u002Fskills\u002Fskill-create\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: skill-create\ndescription: Scaffold a new skill in the monorepo with a valid manifest and catalog entry. Use when creating any new skill.\ndisable-model-invocation: true\nallowed-tools: Read, Grep, Glob, ApplyPatch, Write\nargument-hint: \"[skill-name] [--from \u003Ctool-ref>] [type: opencode|OpenClaw|n8n]\"\n---\n\n# Skill Create — Monorepo Scaffold\n\n## Purpose\nCreate a new skill in the monorepo with a valid `manifest.yaml` and a catalog entry. Enforce secure defaults and prevent duplicates.\n\n## Inputs\n- Skill name (kebab-case)\n- Type: `opencode` | `OpenClaw` | `n8n`\n- Install scope: `local-repo` | `global`\n- Owner\n- Blast radius: `low` | `medium` | `high`\n- Description\n- Permissions (explicit list)\n- Data categories (e.g., `pii`, `financial`, `credentials`, `none`)\n- Egress allowlist (domains or services)\n- Invoke command\u002Ftrigger\n- Phase(s)\n- `--from` flag (optional): A tool reference (CLI command name or URL) to discover capabilities from\n\n- Template paths are resolved from `${AGENTS_SKILLS_ROOT}\u002Fskill-create\u002F`.\n\n## Steps\n1. **Validate naming**\n   - Ensure name is lowercase, kebab-case, and ≤ 64 chars.\n1b. **Discovery phase** (only if `--from` is provided)\n    - Run the discover-tool script:\n      ```bash\n      python3 ${AGENTS_SKILLS_ROOT}\u002Fskill-create\u002Fscripts\u002Fdiscover-tool \"\u003Cfrom-value>\"\n      ```\n    - Parse the JSON output\n    - Use discovered data to pre-populate:\n      - Description (from `description` field)\n      - Dependencies (from `dependencies` field)\n      - Permissions (from `suggested_permissions` field)\n      - Egress (from `suggested_egress` field)\n      - For CLI type: create command docs based on discovered command groups\n      - For API type: create command docs based on discovered endpoints\n    - Present the pre-populated scaffold to the user for review BEFORE writing files\n    - User can modify, remove, or add to any pre-populated field\n    - If user approves, proceed with normal scaffold using pre-populated values\n    - If user rejects or discovery fails, fall back to normal interactive scaffold\n2. **Choose install scope (required)**\n   - Ask the user whether the skill should be installed as:\n     - `local-repo` (Recommended for repo-specific automation) -> `.agents\u002Fskills\u002F\u003Cskill-name>\u002F`\n     - `global` (shared Harnessy skill) -> `tools\u002Fflow-install\u002Fskills\u002F\u003Cskill-name>\u002F`\n   - If the user is unsure, recommend `local-repo` unless the skill is intended for cross-repo reuse through Harnessy installation.\n3. **Check for duplicates (required)**\n   - Search both `.agents\u002Fskills` and `tools\u002Fflow-install\u002Fskills` for matching skill names.\n   - Search `.agents\u002FOpenClaw` and `.agents\u002Fn8n` when applicable.\n   - Search `.jarvis\u002Fcontext\u002Fskills\u002F_catalog.md` for existing entries with the same name or invoke command.\n   - If any match is found, STOP and report the conflict.\n4. **Create directory**\n   - For `local-repo`, create `.agents\u002F\u003Ctype>\u002F\u003Cskill-name>\u002F`.\n   - For `global`, create `tools\u002Fflow-install\u002F\u003Ctype>\u002F\u003Cskill-name>\u002F`.\n5. **Design the intelligence ↔ determinism split**\n   - Put orchestration, intent parsing, and decision-making in `SKILL.md`.\n   - Put repeatable logic (parsing, transforms, formatting, API calls) into `scripts\u002F` inside the skill folder.\n   - If the script should be terminal-callable, name the executable after the final command so installation can expose it through the user-local bin directory (`$XDG_BIN_HOME` or `~\u002F.local\u002Fbin`).\n   - The skill must describe when to call scripts and what input\u002Foutput contracts they expect.\n6. **Write `manifest.yaml`** using `templates\u002Fmanifest.yaml`.\n   - **Autoresearch is mandatory**: all Harnessy skills include `autoresearch: enabled: true` by default. Set `time_budget_seconds` based on blast_radius: high=1800, medium=1200, low=600.\n7. **Write `SKILL.md`** using `templates\u002FSKILL.md` (fill in purpose, inputs, steps, output).\n   - If the skill delegates to command docs, reference them as `${AGENTS_SKILLS_ROOT}\u002F\u003Cskill-name>\u002Fcommands\u002F\u003Cfile-name>.md`.\n   - Do not use fragile references like `.\u002Fcommands\u002F\u003Cfile-name>.md`.\n   - Include this mandatory line in every skill doc: `Template paths are resolved from ${AGENTS_SKILLS_ROOT}\u002F\u003Cskill-name>\u002F.`\n   - When `templates\u002F...` is used, resolve it from `${AGENTS_SKILLS_ROOT}\u002F\u003Cskill-name>\u002F`.\n   - For `local-repo`, keep skill runtime scripts under `.agents\u002Fskills\u002F\u003Cskill-name>\u002Fscripts\u002F`.\n   - For `global`, keep skill runtime scripts under `tools\u002Fflow-install\u002Fskills\u002F\u003Cskill-name>\u002Fscripts\u002F`.\n8. **Add catalog entry** in `.jarvis\u002Fcontext\u002Fskills\u002F_catalog.md` using `templates\u002Fcatalog-entry.md`.\n9. **Register skills when required**\n   - For `local-repo`, run `pnpm skills:register` to install the skill to `${AGENTS_SKILLS_ROOT}`.\n   - For `global`, update the shared source of truth and note any install\u002Fsync step the user should run afterward.\n10. **Return a checklist** for validation and publishing.\n\n## Templates\n- `templates\u002Fmanifest.yaml`\n- `templates\u002FSKILL.md`\n- `templates\u002Fcatalog-entry.md`\n\n## Checklist (Return to User)\n- [ ] Verify `autoresearch` section is present in `manifest.yaml`\n- [ ] Run `\u002Fskill-validate \u003Cskill-name>`\n- [ ] If local-repo: run `pnpm skills:register` (installs to `${AGENTS_SKILLS_ROOT}`)\n- [ ] If global: sync the shared skill install path after review\n- [ ] Add test evidence for blast radius\n- [ ] If high blast radius: get explicit approval before publish\n\n## Deterministic Logic (Scripts)\n- `scripts\u002Fdiscover-tool` — Takes a tool reference string (CLI command or URL), outputs JSON describing the tool's capabilities. Used by the `--from` flag discovery phase.\n  - Input: positional arg (tool reference), optional `--depth N`\n  - Output: JSON to stdout with tool_name, source_type, commands\u002Fendpoints, flags, dependencies\n\n## Feedback Capture\n\nAfter completion, ask the user: **\"Any feedback on this run? (skip to finish)\"**\nIf provided, capture it:\n```bash\npython3 \"${AGENTS_SKILLS_ROOT}\u002F_shared\u002Ftrace_capture.py\" capture \\\n    --skill \"skill-create\" --gate \"run_retrospective\" --gate-type \"retrospective\" \\\n    --outcome \"approved\" --feedback \"\u003Cuser's feedback>\"\n```\n\n","# Skill Create — Monorepo Scaffold\n\n## Purpose\nCreate a new skill in the monorepo with a valid `manifest.yaml` and a catalog entry. Enforce secure defaults and prevent duplicates.\n\n## Inputs\n- Skill name (kebab-case)\n- Type: `opencode` | `OpenClaw` | `n8n`\n- Install scope: `local-repo` | `global`\n- Owner\n- Blast radius: `low` | `medium` | `high`\n- Description\n- Permissions (explicit list)\n- Data categories (e.g., `pii`, `financial`, `credentials`, `none`)\n- Egress allowlist (domains or services)\n- Invoke command\u002Ftrigger\n- Phase(s)\n- `--from` flag (optional): A tool reference (CLI command name or URL) to discover capabilities from\n\n- Template paths are resolved from `${AGENTS_SKILLS_ROOT}\u002Fskill-create\u002F`.\n\n## Steps\n1. **Validate naming**\n   - Ensure name is lowercase, kebab-case, and ≤ 64 chars.\n1b. **Discovery phase** (only if `--from` is provided)\n    - Run the discover-tool script:\n      ```bash\n      python3 ${AGENTS_SKILLS_ROOT}\u002Fskill-create\u002Fscripts\u002Fdiscover-tool \"\u003Cfrom-value>\"\n      ```\n    - Parse the JSON output\n    - Use discovered data to pre-populate:\n      - Description (from `description` field)\n      - Dependencies (from `dependencies` field)\n      - Permissions (from `suggested_permissions` field)\n      - Egress (from `suggested_egress` field)\n      - For CLI type: create command docs based on discovered command groups\n      - For API type: create command docs based on discovered endpoints\n    - Present the pre-populated scaffold to the user for review BEFORE writing files\n    - User can modify, remove, or add to any pre-populated field\n    - If user approves, proceed with normal scaffold using pre-populated values\n    - If user rejects or discovery fails, fall back to normal interactive scaffold\n2. **Choose install scope (required)**\n   - Ask the user whether the skill should be installed as:\n     - `local-repo` (Recommended for repo-specific automation) -> `.agents\u002Fskills\u002F\u003Cskill-name>\u002F`\n     - `global` (shared Harnessy skill) -> `tools\u002Fflow-install\u002Fskills\u002F\u003Cskill-name>\u002F`\n   - If the user is unsure, recommend `local-repo` unless the skill is intended for cross-repo reuse through Harnessy installation.\n3. **Check for duplicates (required)**\n   - Search both `.agents\u002Fskills` and `tools\u002Fflow-install\u002Fskills` for matching skill names.\n   - Search `.agents\u002FOpenClaw` and `.agents\u002Fn8n` when applicable.\n   - Search `.jarvis\u002Fcontext\u002Fskills\u002F_catalog.md` for existing entries with the same name or invoke command.\n   - If any match is found, STOP and report the conflict.\n4. **Create directory**\n   - For `local-repo`, create `.agents\u002F\u003Ctype>\u002F\u003Cskill-name>\u002F`.\n   - For `global`, create `tools\u002Fflow-install\u002F\u003Ctype>\u002F\u003Cskill-name>\u002F`.\n5. **Design the intelligence ↔ determinism split**\n   - Put orchestration, intent parsing, and decision-making in `SKILL.md`.\n   - Put repeatable logic (parsing, transforms, formatting, API calls) into `scripts\u002F` inside the skill folder.\n   - If the script should be terminal-callable, name the executable after the final command so installation can expose it through the user-local bin directory (`$XDG_BIN_HOME` or `~\u002F.local\u002Fbin`).\n   - The skill must describe when to call scripts and what input\u002Foutput contracts they expect.\n6. **Write `manifest.yaml`** using `templates\u002Fmanifest.yaml`.\n   - **Autoresearch is mandatory**: all Harnessy skills include `autoresearch: enabled: true` by default. Set `time_budget_seconds` based on blast_radius: high=1800, medium=1200, low=600.\n7. **Write `SKILL.md`** using `templates\u002FSKILL.md` (fill in purpose, inputs, steps, output).\n   - If the skill delegates to command docs, reference them as `${AGENTS_SKILLS_ROOT}\u002F\u003Cskill-name>\u002Fcommands\u002F\u003Cfile-name>.md`.\n   - Do not use fragile references like `.\u002Fcommands\u002F\u003Cfile-name>.md`.\n   - Include this mandatory line in every skill doc: `Template paths are resolved from ${AGENTS_SKILLS_ROOT}\u002F\u003Cskill-name>\u002F.`\n   - When `templates\u002F...` is used, resolve it from `${AGENTS_SKILLS_ROOT}\u002F\u003Cskill-name>\u002F`.\n   - For `local-repo`, keep skill runtime scripts under `.agents\u002Fskills\u002F\u003Cskill-name>\u002Fscripts\u002F`.\n   - For `global`, keep skill runtime scripts under `tools\u002Fflow-install\u002Fskills\u002F\u003Cskill-name>\u002Fscripts\u002F`.\n8. **Add catalog entry** in `.jarvis\u002Fcontext\u002Fskills\u002F_catalog.md` using `templates\u002Fcatalog-entry.md`.\n9. **Register skills when required**\n   - For `local-repo`, run `pnpm skills:register` to install the skill to `${AGENTS_SKILLS_ROOT}`.\n   - For `global`, update the shared source of truth and note any install\u002Fsync step the user should run afterward.\n10. **Return a checklist** for validation and publishing.\n\n## Templates\n- `templates\u002Fmanifest.yaml`\n- `templates\u002FSKILL.md`\n- `templates\u002Fcatalog-entry.md`\n\n## Checklist (Return to User)\n- [ ] Verify `autoresearch` section is present in `manifest.yaml`\n- [ ] Run `\u002Fskill-validate \u003Cskill-name>`\n- [ ] If local-repo: run `pnpm skills:register` (installs to `${AGENTS_SKILLS_ROOT}`)\n- [ ] If global: sync the shared skill install path after review\n- [ ] Add test evidence for blast radius\n- [ ] If high blast radius: get explicit approval before publish\n\n## Deterministic Logic (Scripts)\n- `scripts\u002Fdiscover-tool` — Takes a tool reference string (CLI command or URL), outputs JSON describing the tool's capabilities. Used by the `--from` flag discovery phase.\n  - Input: positional arg (tool reference), optional `--depth N`\n  - Output: JSON to stdout with tool_name, source_type, commands\u002Fendpoints, flags, dependencies\n\n## Feedback Capture\n\nAfter completion, ask the user: **\"Any feedback on this run? (skip to finish)\"**\nIf provided, capture it:\n```bash\npython3 \"${AGENTS_SKILLS_ROOT}\u002F_shared\u002Ftrace_capture.py\" capture \\\n    --skill \"skill-create\" --gate \"run_retrospective\" --gate-type \"retrospective\" \\\n    --outcome \"approved\" --feedback \"\u003Cuser's feedback>\"\n```\n\n",{"name":5,"description":7,"allowed-tools":24,"argument-hint":25,"disable-model-invocation":26},"Read, Grep, Glob, ApplyPatch, Write","[skill-name] [--from \u003Ctool-ref>] [type: opencode|OpenClaw|n8n]",true,"main","437c399aef1913dab00310feea6d7962d3684949","packages\u002Fcapability-harnessy-v1-full\u002Fresources\u002Fsource\u002Ftools\u002Fflow-install\u002Fskills\u002Fskill-create",6,"https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2","https:\u002F\u002Fgithub.com\u002FFlow-Research","2026-11-12 09:49:54",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":26},[36,39,42],{"tool":37,"command":38},"npx","npx skills add Flow-Research\u002Fharnessy-v2 --skill skill-create",{"tool":40,"command":41},"bunx","bunx skills add Flow-Research\u002Fharnessy-v2 --skill skill-create",{"tool":43,"command":44},"pnpm","pnpm dlx skills add Flow-Research\u002Fharnessy-v2 --skill skill-create","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fcapability-harnessy-v1-full\u002Fresources\u002Fsource\u002Ftools\u002Fflow-install\u002Fskills\u002Fskill-create\u002FSKILL.md",[47,53],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"manifest.yaml","REFERENCE","参考文档",541,"https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fcapability-harnessy-v1-full\u002Fresources\u002Fsource\u002Ftools\u002Fflow-install\u002Fskills\u002Fskill-create\u002Fmanifest.yaml",{"path":54,"resourceType":55,"resourceTypeDesc":56,"sizeBytes":57,"sourceHtmlUrl":58},"scripts\u002Fdiscover-tool","SCRIPT","脚本",15590,"https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fcapability-harnessy-v1-full\u002Fresources\u002Fsource\u002Ftools\u002Fflow-install\u002Fskills\u002Fskill-create\u002Fscripts\u002Fdiscover-tool",[60,67,74,81,88,95],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M77K5PBWD2KS3QRTWR7VHHXB","personal-knowledge","Save sources to the private Harnessy learning library, sync delivered readings, answer questions from saved evidence, browse its wiki, and generate weekly learning reviews.","https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fharnessy-core\u002Fresources\u002Fskills\u002Fpersonal-knowledge\u002FSKILL.md",[],[],{"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},"01M77K5MNRTVCPCES38J7PGJCP","valid-skill","A valid skill for testing purposes.","https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Ftest\u002Ffixtures\u002Fskills\u002Fvalid-skill\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M77K5HQMHXFVQAWSJ9X90BDX","unknown-field","A skill with an unknown frontmatter field.","https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Ftest\u002Ffixtures\u002Fskills\u002Funknown-field\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M77K5GNEGDBJ51MNS14A4N46","nested-child","Nested skill should be ignored.","https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Ftest\u002Ffixtures\u002Fskills\u002Froot-skill-preferred\u002Fnested-child\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M77K5DY7RPPCFCEXV3AFH0R5","root-skill-preferred","Root skill should win.","https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Ftest\u002Ffixtures\u002Fskills\u002Froot-skill-preferred\u002FSKILL.md",[],[],{"skillNo":96,"name":97,"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":98,"categories":99,"keywords":100,"publishedAt":19},"01M77K5CYV52QQ7CWV1TAPG63J","no-frontmatter","https:\u002F\u002Fgithub.com\u002FFlow-Research\u002Fharnessy-v2\u002Fblob\u002Fmain\u002Fpackages\u002Fcoding-agent\u002Ftest\u002Ffixtures\u002Fskills\u002Fno-frontmatter\u002FSKILL.md",[],[]]