[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6GC0X2FN3PM4G16Q4VYPJCX":3,"skill-same:01M6GC0X2FN3PM4G16Q4VYPJCX":110},{"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":29,"publishedAt":50,"linkOnly":51,"markdownContent":52,"markdownBody":53,"frontMatter":54,"refName":88,"commitSha":89,"skillDirectory":90,"forksCount":91,"repositoryHtmlUrl":92,"repositoryHomepageUrl":6,"ownerHtmlUrl":93,"ownerBlogUrl":6,"ownerDescription":94,"updatedAt":95,"install":96,"resources":109},"01M6GC0X2FN3PM4G16Q4VYPJCX","kanban-video-orchestrator",null,"Plan and run multi-agent video production pipelines.","zapabob\u002FSkills","zapabob","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F130830376?v=4","01M6DBMYZAGRTFQDSQYW3W764P","峯岸 亮","01M6DBMZ3GKDSJ01NZ9XTZD7JD",2,"MIT","https:\u002F\u002Fgithub.com\u002Fzapabob\u002FSkills\u002Fblob\u002Fmain\u002Farchive\u002Fskill-md-reviewed\u002Fsha256\u002Fa3\u002Fa30d5a1f4267ceebb054a79a25d985875141bf5e11bf72c36ace5d19daaaf499\u002FSKILL.md",[18,23,26],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000015","视觉设计","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000016","图像与多媒体",{"categoryNo":27,"categoryName":28,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发",[30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49],"kanban video orchestrator","Video","orchestrator","kanban","diagrams","pixel-art","SOUL.md","project","pixel","soul","infographic","Art","design","model","team","pipeline","search","humanizer","architecture","orchestration","2026-11-10 10:12:30",false,"---\nname: kanban-video-orchestrator\ndescription: Plan and run multi-agent video production pipelines.\nversion: 1.0.0\nauthor: [SHL0MS, alt-glitch]\nlicense: MIT\nplatforms: [linux, macos, windows]\nmetadata:\n  hermes:\n    tags: [video, kanban, multi-agent, orchestration, production-pipeline]\n    related_skills: [ascii-video, manim-video, p5js, comfyui, pixel-art, ascii-art, songwriting-and-ai-music, heartmula, songsee, youtube-content, claude-design, excalidraw, architecture-diagram, concept-diagrams, baoyu-comic, baoyu-infographic, humanizer, gif-search, meme-generation]\n    credits: |\n      The single-project workspace layout, profile-config patching pattern,\n      SOUL.md-per-profile model, TEAM.md task-graph convention, and\n      `--workspace dir:\u002Fabs\u002Fpath` discipline are adapted from alt-glitch's\n      original multi-agent video pipeline at\n      https:\u002F\u002Fgithub.com\u002FNousResearch\u002Fkanban-video-pipeline.\n---\n\n# Kanban Video Orchestrator\n\nWrap any video request — from a 15-second product teaser to a 5-minute narrative\nshort to a music video to an ASCII loop — in a Hermes Kanban pipeline that\ndecomposes the work to specialized agent profiles.\n\nThis skill does **not** render anything itself. It is a meta-pipeline that:\n\n1. **Scopes** the request through targeted discovery\n2. **Designs** an appropriate team (which roles, which tools per role) based on the style\n3. **Generates** a setup script that creates Hermes profiles, project workspace, and the initial kanban task\n4. **Hands off** to the director profile, which decomposes via the kanban\n5. **Monitors** execution, helps intervene when tasks stall or fail\n\nThe actual rendering happens inside the kanban once it's running, via whichever\nexisting skills + tools fit the scenes — `ascii-video`, `manim-video`, `p5js`,\n`comfyui`, `touchdesigner-mcp`, `songwriting-and-ai-music`,\n`heartmula`, external APIs, or plain Python with PIL + ffmpeg.\n\n## When NOT to use this skill\n\n- The video is one continuous procedural project that needs no specialists. Just write the code directly.\n- The user wants a quick one-shot conversion (e.g. \"convert this mp4 to a GIF\") — use ffmpeg directly.\n- The output is a static image, GIF, or audio-only artifact — use the matching specific skill (`ascii-art`, `gifs`, `meme-generation`, `songwriting-and-ai-music`).\n- The work fits a single existing skill cleanly (e.g. a pure ASCII video — just use `ascii-video`).\n\n## Workflow\n\n```\nDISCOVER  →  BRIEF  →  TEAM DESIGN  →  SETUP  →  EXECUTE  →  MONITOR\n```\n\n### Step 1 — Discover (ask the right questions)\n\nThe discovery process is **adaptive**: ask only what is actually needed. Always\nstart with three questions to identify the broad shape:\n\n- **What is the video?** (one-sentence brief)\n- **How long?** (5-30s teaser \u002F 30-90s short \u002F 90s-3min explainer \u002F 3-10min film \u002F longer)\n- **What aspect ratio + target platform?** (1:1 \u002F 9:16 \u002F 16:9; X, IG, YouTube, internal, etc.)\n\nFrom the answer, classify the style category. The style determines which\nfollow-up questions to ask. **Do not ask all questions at once.** Ask 2-4 at a\ntime, listen, then proceed. Make reasonable assumptions whenever the user\nimplies an answer.\n\nFor complete intake patterns and per-style question banks, see\n**[references\u002Fintake.md](references\u002Fintake.md)**.\n\n### Step 2 — Brief\n\nOnce enough is known, produce a structured `brief.md` using the template in\n`assets\u002Fbrief.md.tmpl`. Stages:\n\n1. **Concept** — the one-sentence pitch + emotional north star\n2. **Scope** — duration, aspect, platform, deadline\n3. **Style** — visual references, brand constraints, tone\n4. **Scenes** — beat-by-beat breakdown (durations, content, target tool)\n5. **Audio** — narration \u002F music \u002F SFX \u002F silent (per scene if needed)\n6. **Deliverables** — file format, resolution, optional alternates (vertical cut, GIF, etc.)\n\nShow the brief to the user for confirmation before designing the team. **The\nbrief is the contract** — every downstream task references it.\n\n### Step 3 — Team design\n\nPick role archetypes from the library that fit this video. **Compose, don't\nclone.** Most videos need 4-7 profiles. The director is always present; the\nrest are picked by what the brief actually requires.\n\nFor the role library and per-style team compositions, see\n**[references\u002Frole-archetypes.md](references\u002Frole-archetypes.md)**.\n\nFor mapping role → which Hermes skills + toolsets it loads, see\n**[references\u002Ftool-matrix.md](references\u002Ftool-matrix.md)**.\n\n### Step 4 — Setup\n\nGenerate a setup script (`setup.sh`) and run it. The script:\n\n1. Creates the project workspace (`~\u002Fprojects\u002Fvideo-pipeline\u002F\u003Cslug>\u002F`)\n2. Copies any provided assets into `taste\u002F`, `audio\u002F`, `assets\u002F`\n3. Creates each Hermes profile via `hermes profile create --clone`\n4. Writes per-profile `SOUL.md` (personality + role definition)\n5. Configures profile YAML (toolsets, always_load skills, cwd)\n6. Writes `brief.md`, `TEAM.md`, and `taste\u002F` content\n7. Fires the initial `hermes kanban create` task assigned to the director\n\nUse `scripts\u002Fbootstrap_pipeline.py` to generate setup.sh from a brief +\nteam-design JSON. See **[references\u002Fkanban-setup.md](references\u002Fkanban-setup.md)**\nfor the setup script structure, profile config patterns, and the critical\n\"shared workspace\" rule.\n\n### Step 5 — Execute\n\nRun `setup.sh`. Then provide the user with monitoring commands:\n\n```bash\nhermes kanban watch --tenant \u003Cproject-tenant>     # live events\nhermes kanban list  --tenant \u003Cproject-tenant>     # board snapshot\nhermes dashboard                                   # visual board UI\n```\n\nThe director profile takes over from here, decomposing the work and routing\ntasks to specialist profiles via the kanban toolset.\n\n### Step 6 — Monitor and intervene\n\nStay engaged — the kanban runs autonomously but a stuck task or bad output\nneeds human (or AI) judgment.\n\nMonitoring patterns: poll `kanban list` periodically, inspect any RUNNING task\nthat exceeds its expected duration with `kanban show \u003Cid>`, and check\nheartbeats. When a worker's output fails review, the standard interventions are:\n\n1. Comment on the worker's task with specific feedback (`kanban_comment`)\n2. Create a re-run task with the original as parent\n3. Adjust the brief's scope and let the director re-decompose\n\nFor diagnostic patterns, intervention recipes, and the \"task is stuck\"\nplaybook, see **[references\u002Fmonitoring.md](references\u002Fmonitoring.md)**.\n\n## Reference: worked examples\n\nSix concrete pipelines covering very different video styles — narrative film,\nproduct\u002Fmarketing, music video, math\u002Falgorithm explainer, ASCII video, real-time\ninstallation — showing how the same workflow yields very different teams and\ntask graphs. See **[references\u002Fexamples.md](references\u002Fexamples.md)**.\n\n## Critical rules\n\n1. **Discovery before action.** Never start generating a brief or team without\n   asking at least the three baseline questions. A bad brief cascades through\n   the entire pipeline.\n\n2. **Match the team to the video.** Don't reuse the same 4-profile setup for\n   every job. A music video that doesn't have a beat-analysis profile will\n   misfire. A narrative film that doesn't have a writer profile will produce\n   incoherent scenes. See `references\u002Frole-archetypes.md`.\n\n3. **One workspace per project.** All profiles for a given video share the same\n   `dir:` workspace. Tasks pass artifacts via shared filesystem and structured\n   handoffs. **Every** `kanban_create` call passes\n   `workspace_kind=\"dir\"` + `workspace_path=\"\u003Cabsolute project path>\"`.\n\n4. **Tenant every project.** Use a project-specific tenant\n   (`--tenant \u003Cproject-slug>`). Keeps the dashboard scoped and prevents\n   cross-pollination with other ongoing kanbans.\n\n5. **Respect existing skills.** When a scene fits an existing skill, the\n   relevant renderer should load that skill via `--skill \u003Cname>` on its task\n   or `always_load` in its profile. Do not re-derive what a skill already\n   provides.\n\n6. **The director never executes.** Even with the full `kanban + terminal +\n   file` toolset, the director's `SOUL.md` rules forbid it from executing\n   work itself. It decomposes and routes only — every concrete task becomes\n   a `hermes kanban create` call to a specialist profile. The kanban\n   orchestration guidance auto-injected into every kanban worker's system\n   prompt spells this out further.\n\n7. **Don't over-decompose.** A 30-second product video does NOT need 20 tasks.\n   Aim for the smallest task graph that still parallelizes well and exposes the\n   right human-review gates.\n\n8. **Verify API keys BEFORE firing.** External APIs (TTS, image-gen,\n   image-to-video) need keys in `${HERMES_HOME:-~\u002F.hermes}\u002F.env` or the user's secret store.\n   A worker that hits a missing-key error wastes a task slot. The setup\n   script's `check_key` helper aborts cleanly if a required key is missing.\n\n## File map\n\n```\nSKILL.md                            ← this file (workflow + rules)\nreferences\u002F\n  intake.md                         ← discovery question banks per style\n  role-archetypes.md                ← role library (writer, designer, animator, …)\n  tool-matrix.md                    ← skill + toolset mapping per role\n  kanban-setup.md                   ← setup script structure & profile config\n  monitoring.md                     ← watch + intervene patterns\n  examples.md                       ← six worked pipelines\nassets\u002F\n  brief.md.tmpl                     ← brief skeleton\n  setup.sh.tmpl                     ← setup script skeleton\n  soul.md.tmpl                      ← profile personality skeleton\nscripts\u002F\n  bootstrap_pipeline.py             ← generate setup.sh from brief + team JSON\n  monitor.py                        ← polling + intervention helpers\n```\n","# Kanban Video Orchestrator\n\nWrap any video request — from a 15-second product teaser to a 5-minute narrative\nshort to a music video to an ASCII loop — in a Hermes Kanban pipeline that\ndecomposes the work to specialized agent profiles.\n\nThis skill does **not** render anything itself. It is a meta-pipeline that:\n\n1. **Scopes** the request through targeted discovery\n2. **Designs** an appropriate team (which roles, which tools per role) based on the style\n3. **Generates** a setup script that creates Hermes profiles, project workspace, and the initial kanban task\n4. **Hands off** to the director profile, which decomposes via the kanban\n5. **Monitors** execution, helps intervene when tasks stall or fail\n\nThe actual rendering happens inside the kanban once it's running, via whichever\nexisting skills + tools fit the scenes — `ascii-video`, `manim-video`, `p5js`,\n`comfyui`, `touchdesigner-mcp`, `songwriting-and-ai-music`,\n`heartmula`, external APIs, or plain Python with PIL + ffmpeg.\n\n## When NOT to use this skill\n\n- The video is one continuous procedural project that needs no specialists. Just write the code directly.\n- The user wants a quick one-shot conversion (e.g. \"convert this mp4 to a GIF\") — use ffmpeg directly.\n- The output is a static image, GIF, or audio-only artifact — use the matching specific skill (`ascii-art`, `gifs`, `meme-generation`, `songwriting-and-ai-music`).\n- The work fits a single existing skill cleanly (e.g. a pure ASCII video — just use `ascii-video`).\n\n## Workflow\n\n```\nDISCOVER  →  BRIEF  →  TEAM DESIGN  →  SETUP  →  EXECUTE  →  MONITOR\n```\n\n### Step 1 — Discover (ask the right questions)\n\nThe discovery process is **adaptive**: ask only what is actually needed. Always\nstart with three questions to identify the broad shape:\n\n- **What is the video?** (one-sentence brief)\n- **How long?** (5-30s teaser \u002F 30-90s short \u002F 90s-3min explainer \u002F 3-10min film \u002F longer)\n- **What aspect ratio + target platform?** (1:1 \u002F 9:16 \u002F 16:9; X, IG, YouTube, internal, etc.)\n\nFrom the answer, classify the style category. The style determines which\nfollow-up questions to ask. **Do not ask all questions at once.** Ask 2-4 at a\ntime, listen, then proceed. Make reasonable assumptions whenever the user\nimplies an answer.\n\nFor complete intake patterns and per-style question banks, see\n**[references\u002Fintake.md](references\u002Fintake.md)**.\n\n### Step 2 — Brief\n\nOnce enough is known, produce a structured `brief.md` using the template in\n`assets\u002Fbrief.md.tmpl`. Stages:\n\n1. **Concept** — the one-sentence pitch + emotional north star\n2. **Scope** — duration, aspect, platform, deadline\n3. **Style** — visual references, brand constraints, tone\n4. **Scenes** — beat-by-beat breakdown (durations, content, target tool)\n5. **Audio** — narration \u002F music \u002F SFX \u002F silent (per scene if needed)\n6. **Deliverables** — file format, resolution, optional alternates (vertical cut, GIF, etc.)\n\nShow the brief to the user for confirmation before designing the team. **The\nbrief is the contract** — every downstream task references it.\n\n### Step 3 — Team design\n\nPick role archetypes from the library that fit this video. **Compose, don't\nclone.** Most videos need 4-7 profiles. The director is always present; the\nrest are picked by what the brief actually requires.\n\nFor the role library and per-style team compositions, see\n**[references\u002Frole-archetypes.md](references\u002Frole-archetypes.md)**.\n\nFor mapping role → which Hermes skills + toolsets it loads, see\n**[references\u002Ftool-matrix.md](references\u002Ftool-matrix.md)**.\n\n### Step 4 — Setup\n\nGenerate a setup script (`setup.sh`) and run it. The script:\n\n1. Creates the project workspace (`~\u002Fprojects\u002Fvideo-pipeline\u002F\u003Cslug>\u002F`)\n2. Copies any provided assets into `taste\u002F`, `audio\u002F`, `assets\u002F`\n3. Creates each Hermes profile via `hermes profile create --clone`\n4. Writes per-profile `SOUL.md` (personality + role definition)\n5. Configures profile YAML (toolsets, always_load skills, cwd)\n6. Writes `brief.md`, `TEAM.md`, and `taste\u002F` content\n7. Fires the initial `hermes kanban create` task assigned to the director\n\nUse `scripts\u002Fbootstrap_pipeline.py` to generate setup.sh from a brief +\nteam-design JSON. See **[references\u002Fkanban-setup.md](references\u002Fkanban-setup.md)**\nfor the setup script structure, profile config patterns, and the critical\n\"shared workspace\" rule.\n\n### Step 5 — Execute\n\nRun `setup.sh`. Then provide the user with monitoring commands:\n\n```bash\nhermes kanban watch --tenant \u003Cproject-tenant>     # live events\nhermes kanban list  --tenant \u003Cproject-tenant>     # board snapshot\nhermes dashboard                                   # visual board UI\n```\n\nThe director profile takes over from here, decomposing the work and routing\ntasks to specialist profiles via the kanban toolset.\n\n### Step 6 — Monitor and intervene\n\nStay engaged — the kanban runs autonomously but a stuck task or bad output\nneeds human (or AI) judgment.\n\nMonitoring patterns: poll `kanban list` periodically, inspect any RUNNING task\nthat exceeds its expected duration with `kanban show \u003Cid>`, and check\nheartbeats. When a worker's output fails review, the standard interventions are:\n\n1. Comment on the worker's task with specific feedback (`kanban_comment`)\n2. Create a re-run task with the original as parent\n3. Adjust the brief's scope and let the director re-decompose\n\nFor diagnostic patterns, intervention recipes, and the \"task is stuck\"\nplaybook, see **[references\u002Fmonitoring.md](references\u002Fmonitoring.md)**.\n\n## Reference: worked examples\n\nSix concrete pipelines covering very different video styles — narrative film,\nproduct\u002Fmarketing, music video, math\u002Falgorithm explainer, ASCII video, real-time\ninstallation — showing how the same workflow yields very different teams and\ntask graphs. See **[references\u002Fexamples.md](references\u002Fexamples.md)**.\n\n## Critical rules\n\n1. **Discovery before action.** Never start generating a brief or team without\n   asking at least the three baseline questions. A bad brief cascades through\n   the entire pipeline.\n\n2. **Match the team to the video.** Don't reuse the same 4-profile setup for\n   every job. A music video that doesn't have a beat-analysis profile will\n   misfire. A narrative film that doesn't have a writer profile will produce\n   incoherent scenes. See `references\u002Frole-archetypes.md`.\n\n3. **One workspace per project.** All profiles for a given video share the same\n   `dir:` workspace. Tasks pass artifacts via shared filesystem and structured\n   handoffs. **Every** `kanban_create` call passes\n   `workspace_kind=\"dir\"` + `workspace_path=\"\u003Cabsolute project path>\"`.\n\n4. **Tenant every project.** Use a project-specific tenant\n   (`--tenant \u003Cproject-slug>`). Keeps the dashboard scoped and prevents\n   cross-pollination with other ongoing kanbans.\n\n5. **Respect existing skills.** When a scene fits an existing skill, the\n   relevant renderer should load that skill via `--skill \u003Cname>` on its task\n   or `always_load` in its profile. Do not re-derive what a skill already\n   provides.\n\n6. **The director never executes.** Even with the full `kanban + terminal +\n   file` toolset, the director's `SOUL.md` rules forbid it from executing\n   work itself. It decomposes and routes only — every concrete task becomes\n   a `hermes kanban create` call to a specialist profile. The kanban\n   orchestration guidance auto-injected into every kanban worker's system\n   prompt spells this out further.\n\n7. **Don't over-decompose.** A 30-second product video does NOT need 20 tasks.\n   Aim for the smallest task graph that still parallelizes well and exposes the\n   right human-review gates.\n\n8. **Verify API keys BEFORE firing.** External APIs (TTS, image-gen,\n   image-to-video) need keys in `${HERMES_HOME:-~\u002F.hermes}\u002F.env` or the user's secret store.\n   A worker that hits a missing-key error wastes a task slot. The setup\n   script's `check_key` helper aborts cleanly if a required key is missing.\n\n## File map\n\n```\nSKILL.md                            ← this file (workflow + rules)\nreferences\u002F\n  intake.md                         ← discovery question banks per style\n  role-archetypes.md                ← role library (writer, designer, animator, …)\n  tool-matrix.md                    ← skill + toolset mapping per role\n  kanban-setup.md                   ← setup script structure & profile config\n  monitoring.md                     ← watch + intervene patterns\n  examples.md                       ← six worked pipelines\nassets\u002F\n  brief.md.tmpl                     ← brief skeleton\n  setup.sh.tmpl                     ← setup script skeleton\n  soul.md.tmpl                      ← profile personality skeleton\nscripts\u002F\n  bootstrap_pipeline.py             ← generate setup.sh from brief + team JSON\n  monitor.py                        ← polling + intervention helpers\n```\n",{"name":5,"author":55,"license":15,"version":58,"metadata":59,"platforms":84,"description":7},[56,57],"SHL0MS","alt-glitch","1.0.0",{"hermes":60},{"tags":61,"credits":65,"related_skills":66},[62,33,63,49,64],"video","multi-agent","production-pipeline","The single-project workspace layout, profile-config patching pattern,\nSOUL.md-per-profile model, TEAM.md task-graph convention, and\n`--workspace dir:\u002Fabs\u002Fpath` discipline are adapted from alt-glitch's\noriginal multi-agent video pipeline at\nhttps:\u002F\u002Fgithub.com\u002FNousResearch\u002Fkanban-video-pipeline.",[67,68,69,70,35,71,72,73,74,75,76,77,78,79,80,81,47,82,83],"ascii-video","manim-video","p5js","comfyui","ascii-art","songwriting-and-ai-music","heartmula","songsee","youtube-content","claude-design","excalidraw","architecture-diagram","concept-diagrams","baoyu-comic","baoyu-infographic","gif-search","meme-generation",[85,86,87],"linux","macos","windows","main","0d1434f89f4bc6b07e9d272accd8c1013768820f","archive\u002Fskill-md-reviewed\u002Fsha256\u002Fa3\u002Fa30d5a1f4267ceebb054a79a25d985875141bf5e11bf72c36ace5d19daaaf499",0,"https:\u002F\u002Fgithub.com\u002Fzapabob\u002FSkills","https:\u002F\u002Fgithub.com\u002Fzapabob","The Open University of JAPAN　2019-now\r\n\r\nTokyo University of Science　2007-2010","2026-10-26 16:29:01",{"source":8,"skillName":5,"commands":97,"prompt":107,"zipAvailable":108},[98,101,104],{"tool":99,"command":100},"npx","npx skills add zapabob\u002FSkills --skill kanban-video-orchestrator",{"tool":102,"command":103},"bunx","bunx skills add zapabob\u002FSkills --skill kanban-video-orchestrator",{"tool":105,"command":106},"pnpm","pnpm dlx skills add zapabob\u002FSkills --skill kanban-video-orchestrator","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fzapabob\u002FSkills\u002Fblob\u002Fmain\u002Farchive\u002Fskill-md-reviewed\u002Fsha256\u002Fa3\u002Fa30d5a1f4267ceebb054a79a25d985875141bf5e11bf72c36ace5d19daaaf499\u002FSKILL.md",true,[],[111,141,173,197,224,241],{"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":126,"publishedAt":50},"01M6GJTE1GESGGF9K028Y7753E","xurl","X\u002FTwitter via xurl CLI: post, search, DM, media, v2 API.","https:\u002F\u002Fgithub.com\u002Fzapabob\u002FSkills\u002Fblob\u002Fmain\u002Fregistry\u002Fskills\u002Fxurl\u002Fvariants\u002Fhermes-agent\u002FSKILL.md",[117,120,121],{"categoryNo":118,"categoryName":119,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000010","后端与接口",{"categoryNo":27,"categoryName":28,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":122,"categoryName":123,"taxonomyType":124,"taxonomyTypeDesc":125},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[113,127,86,128,129,85,87,130,46,131,132,133,134,135,136,137,88,138,139,140],"openclaw","cli","github","official","https","social-media","api","Media","hermes","hermes agent","twitter","social","via","post",{"skillNo":142,"name":143,"nameZh":6,"description":144,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":145,"categories":146,"keywords":156,"publishedAt":50},"01M6GJT8S1PXQ8WPXT1HSNBWH2","writing-plans","Write implementation plans: bite-sized tasks, paths, code.","https:\u002F\u002Fgithub.com\u002Fzapabob\u002FSkills\u002Fblob\u002Fmain\u002Fregistry\u002Fskills\u002Fwriting-plans\u002Fvariants\u002Fhermes-agent\u002FSKILL.md",[147,148,151,152,153],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":149,"categoryName":150,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":27,"categoryName":28,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":122,"categoryName":123,"taxonomyType":124,"taxonomyTypeDesc":125},{"categoryNo":154,"categoryName":155,"taxonomyType":124,"taxonomyTypeDesc":125},"01JBSEEDCAT000000000000036","内容创作者",[157,158,159,42,86,160,85,161,87,162,163,164,165,166,167,168,169,170,171,172],"writing","writing plans","plans","review","workflow","documentation","tasks","development","code","planning","code-review","subagent","write","superpowers","paths","test-driven-development",{"skillNo":174,"name":175,"nameZh":6,"description":176,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":177,"categories":178,"keywords":180,"publishedAt":50},"01M6GJST0H5X7WHBJ3BZK0DVKH","whisper","OpenAI's general-purpose speech recognition model. Supports 99 languages, transcription, translation to English, and language identification. Six model sizes from tiny (39M params) to large (1550M params). Use for speech-to-text, podcast transcription, or multilingual audio processing. Best for robust, multilingual ASR.","https:\u002F\u002Fgithub.com\u002Fzapabob\u002FSkills\u002Fblob\u002Fmain\u002Fregistry\u002Fskills\u002Fwhisper\u002Fvariants\u002Fhermes-agent-optional\u002FSKILL.md",[179],{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},[175,86,85,181,182,183,184,185,186,187,43,188,189,190,191,192,193,194,195,196],"openai","speech","multimodal","Transformers","ASR","Audio","text","speech recognition","use","general","multilingual","translation","english","identification","processing","Audio Processing",{"skillNo":198,"name":199,"nameZh":6,"description":200,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":201,"categories":202,"keywords":206,"publishedAt":50},"01M6GJSMRQ8EETZTYMG72P70EA","weights-and-biases","W&B: log ML experiments, sweeps, model registry, dashboards.","https:\u002F\u002Fgithub.com\u002Fzapabob\u002FSkills\u002Fblob\u002Fmain\u002Fregistry\u002Fskills\u002Fweights-and-biases\u002Fvariants\u002Fhermes-agent\u002FSKILL.md",[203],{"categoryNo":204,"categoryName":205,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000013","分析与可视化",[207,86,208,209,87,210,211,212,213,214,215,43,216,217,218,219,220,221,222,223],"weights and biases","tensorflow","experiments","pytorch","visualization","registry","real-time","tuning","dashboards","experiment","log","time","Experiment Tracking","Model Registry","Hyperparameter Tuning","WandB","Real-Time Visualization",{"skillNo":225,"name":226,"nameZh":6,"description":227,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":228,"categories":229,"keywords":231,"publishedAt":50},"01M6GJS19KDPRY112JTR326S06","watchers","Poll RSS, JSON APIs, and GitHub with watermark dedup.","https:\u002F\u002Fgithub.com\u002Fzapabob\u002FSkills\u002Fblob\u002Fmain\u002Fregistry\u002Fskills\u002Fwatchers\u002Fvariants\u002Fhermes-agent-optional\u002FSKILL.md",[230],{"categoryNo":27,"categoryName":28,"taxonomyType":21,"taxonomyTypeDesc":22},[226,86,232,129,85,233,234,235,236,237,238,135,136,239,240],"terminal","automation","http","monitoring","cron","json","rss","devops","with",{"skillNo":242,"name":243,"nameZh":6,"description":244,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":245,"categories":246,"keywords":249,"publishedAt":50},"01M6GJRHJD2A6Q830SWF9W48BC","serving-llms-vllm","vLLM: high-throughput LLM serving, OpenAI API, quantization.","https:\u002F\u002Fgithub.com\u002Fzapabob\u002FSkills\u002Fblob\u002Fmain\u002Fregistry\u002Fskills\u002Fvllm\u002Fvariants\u002Fhermes-agent\u002FSKILL.md",[247,248],{"categoryNo":118,"categoryName":119,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":122,"categoryName":123,"taxonomyType":124,"taxonomyTypeDesc":125},[250,251,252,85,253,181,254,255,256,184,257,133,258,259,260,261,262,263,264,265],"llms","vLLM","serving llms vllm","research","openai api","tensor","Inference Serving","Production","llm serving","High Throughput","PagedAttention","Continuous Batching","Quantization","Tensor Parallelism","high-throughput","throughput"]