[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7344WSFXSY3VCT22AWRJS1F":3,"skill-same:01M7344WSFXSY3VCT22AWRJS1F":107},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":15,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":34,"install":35,"resources":48},"01M7344WSFXSY3VCT22AWRJS1F","higgsfield-generate",null,"Generate images\u002Fvideos\u002F3D assets\u002Faudio via Higgsfield AI. Defaults:\nGPT Image 2 for image\u002Fdesign\u002Ftext, Seedance 2.0 for\nvideo, Nano Banana 2\u002FPro for character\u002Freference images,\nMarketing Studio for ads, Sonilo\u002FMirelo for audio, plus\nSoul models and Kling 3.0.\nUse when: \"generate an image\", \"make a video\", \"animate\nthis photo\", \"image-to-video\", \"edit\u002Fstylize\u002Fremix this\nimage\", \"produce a clip\", \"reframe this video\", \"edit\nthis video from a sketch\", \"create a 3D model\", \"make a\nGLB\u002Fmesh\", \"create a sound effect\", \"make music\",\n\"text-to-audio\", \"create an ad\", \"make a UGC video\",\n\"product demo\", \"unboxing\", \"brand video\", \"presenter\nvideo\", \"import product from URL\", \"create avatar for ad\",\nor \"analyze video virality\". Supports image-to-image,\nimage-to-video, image-to-3D (`multi_image_to_3d`),\ntext-to-audio (`mirelo_text_to_audio`), text-to-music\n(`sonilo_music`), workflow generation (`draw_to_video`,\n`reframe`), references, job\u002Fupload IDs, Marketing Studio,\nand Virality Predictor (`brain_activity`).\nSupports two execution surfaces: Higgsfield Cloud API\nwith HIGGSFIELD_API_KEY_ID\u002FHIGGSFIELD_API_KEY_SECRET,\nand the standard consumer Higgsfield CLI\u002Fdashboard.\nChain with higgsfield-soul-id for face\u002Fidentity consistency.\nNOT for: Soul Character training (use higgsfield-soul-id),\nproduct photoshoots, marketplace listing cards,\ntext\u002Fchat\u002FTTS tasks.","LiamVisionary\u002Fhivemindos","LiamVisionary","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F102138294?v=4","01M73379NZZ99TT4ZS17XR3Q9D","01M73379PX6AKQNWN05VNSYRXW",13,"MIT","https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002FLICENSE","https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nversion: 0.3.0\nname: higgsfield-generate\ndescription: |\n  Generate images\u002Fvideos\u002F3D assets\u002Faudio via Higgsfield AI. Defaults:\n  GPT Image 2 for image\u002Fdesign\u002Ftext, Seedance 2.0 for\n  video, Nano Banana 2\u002FPro for character\u002Freference images,\n  Marketing Studio for ads, Sonilo\u002FMirelo for audio, plus\n  Soul models and Kling 3.0.\n  Use when: \"generate an image\", \"make a video\", \"animate\n  this photo\", \"image-to-video\", \"edit\u002Fstylize\u002Fremix this\n  image\", \"produce a clip\", \"reframe this video\", \"edit\n  this video from a sketch\", \"create a 3D model\", \"make a\n  GLB\u002Fmesh\", \"create a sound effect\", \"make music\",\n  \"text-to-audio\", \"create an ad\", \"make a UGC video\",\n  \"product demo\", \"unboxing\", \"brand video\", \"presenter\n  video\", \"import product from URL\", \"create avatar for ad\",\n  or \"analyze video virality\". Supports image-to-image,\n  image-to-video, image-to-3D (`multi_image_to_3d`),\n  text-to-audio (`mirelo_text_to_audio`), text-to-music\n  (`sonilo_music`), workflow generation (`draw_to_video`,\n  `reframe`), references, job\u002Fupload IDs, Marketing Studio,\n  and Virality Predictor (`brain_activity`).\n  Supports two execution surfaces: Higgsfield Cloud API\n  with HIGGSFIELD_API_KEY_ID\u002FHIGGSFIELD_API_KEY_SECRET,\n  and the standard consumer Higgsfield CLI\u002Fdashboard.\n  Chain with higgsfield-soul-id for face\u002Fidentity consistency.\n  NOT for: Soul Character training (use higgsfield-soul-id),\n  product photoshoots, marketplace listing cards,\n  text\u002Fchat\u002FTTS tasks.\nargument-hint: \"[prompt-or-analysis-request] [--model \u003Cname>] [--image|--video \u003Cpath-or-id>]\"\nallowed-tools: Bash\n---\n\n# Higgsfield Generate\n\nSubmit jobs to Higgsfield models through the right surface for the user's intent. Supports Higgsfield Cloud API jobs and the standard consumer `higgsfield` CLI\u002Fdashboard. Covers generic image\u002Fvideo\u002F3D\u002Faudio generation, Marketing Studio (branded ads, avatars, products, hooks, settings), and, secondarily, Virality Predictor video scoring.\n\n## Surface Selection - Cloud API vs Consumer CLI\n\nChoose the execution surface before any auth, install, upload, or generation step.\n\n1. If the user says \"Cloud\", \"Higgsfield Cloud\", \"API\", \"API key\", \"key id\", \"secret\", \"server-side\", \"use my env\", or \"no dashboard login\", use Higgsfield Cloud API.\n2. If the user says \"consumer\", \"standard\", \"regular Higgsfield\", \"dashboard\", \"CLI\", \"use my logged-in Higgsfield account\", or explicitly asks for the local `higgsfield` command, use the consumer CLI\u002Fdashboard.\n3. If the user has not specified and both surfaces could satisfy the request, ask one short question before proceeding: \"Use Higgsfield Cloud API or consumer Higgsfield CLI?\"\n4. Never silently switch surfaces. In particular, do not run `higgsfield auth login` for a Cloud\u002FAPI-key request, and do not fall back to procedural\u002Flocal generation when a Higgsfield generation was requested.\n\n### Cloud API execution\n\n1. Verify credentials by name only:\n   ```bash\n   hive-env-check HIGGSFIELD_API_KEY_ID\n   hive-env-check HIGGSFIELD_API_KEY_SECRET\n   ```\n2. Discover docs from the official LLM index before exploring further:\n   ```text\n   https:\u002F\u002Fdocs.higgsfield.ai\u002Fdocs\u002Fllms.txt\n   ```\n3. Use the documented Cloud REST contract:\n   - Base URL: `https:\u002F\u002Fplatform.higgsfield.ai`\n   - Submit: `POST https:\u002F\u002Fplatform.higgsfield.ai\u002F{model_id}`\n   - Status: `GET https:\u002F\u002Fplatform.higgsfield.ai\u002Frequests\u002F{request_id}\u002Fstatus`\n   - Cancel queued request: `POST https:\u002F\u002Fplatform.higgsfield.ai\u002Frequests\u002F{request_id}\u002Fcancel`\n   - Webhook: append `?hf_webhook=\u003Chttps-url>` to the model submit URL.\n   - Auth header: `Authorization: Key \u003Capi_key>:\u003Capi_secret>`\n4. Run Cloud API calls through the shared env loader and keep credentials in env expansion, not logs:\n   ```bash\n   hive-env-run -- sh -c 'curl -sS \\\n     -H \"Authorization: Key ${HIGGSFIELD_API_KEY_ID}:${HIGGSFIELD_API_KEY_SECRET}\" \\\n     -H \"Content-Type: application\u002Fjson\" \\\n     -H \"Accept: application\u002Fjson\" \\\n     -d @payload.json \\\n     https:\u002F\u002Fplatform.higgsfield.ai\u002F\u003Cmodel_id>'\n   ```\n5. Do not print, copy, log, or persist secret values. Redact token\u002Fkey\u002Fsecret fields before showing any JSON output.\n6. Prefer `curl` for direct REST checks. Python `urllib` can be blocked by Cloudflare Error 1010 (`browser_signature_banned`) before auth is evaluated. A safe no-cost auth\u002Fpath sanity check is a fake request-id status lookup; valid platform routing should return `404 {\"detail\":\"Not found\"}`, not `Missing token`, `Invalid token`, or consumer workspace credit errors:\n   ```bash\n   hive-env-run -- sh -c 'curl -sS \\\n     -H \"Accept: application\u002Fjson\" \\\n     -H \"Authorization: Key ${HIGGSFIELD_API_KEY_ID}:${HIGGSFIELD_API_KEY_SECRET}\" \\\n     https:\u002F\u002Fplatform.higgsfield.ai\u002Frequests\u002F00000000-0000-0000-0000-000000000000\u002Fstatus'\n   ```\n7. The official Python SDK is `higgsfield-client`. It accepts either `HF_KEY=\u003Capi_key>:\u003Capi_secret>` or `HF_API_KEY` plus `HF_API_SECRET`; map from the HivemindOS shared env at process launch rather than writing new secret files.\n8. Use the same model-selection guidance below, then translate prompt\u002Fmedia\u002Fmodel\u002Fduration\u002Fresolution\u002Faspect choices into the Cloud API submit\u002Fpoll\u002Fdownload lifecycle.\n9. For Seedance 2.0, Kling 3.0, or any model with known non-standard behavior, read `higgsfield-api-quirks` before building the payload. In particular, Seedance 2.0 audio must be sent as root-level `input_audio`, not as a `role: \"audio\"` item in `medias`, and aspect ratio must be the string `aspect_ratio`.\n\nCloud model examples from the official docs:\n\n- Text-to-image: `higgsfield-ai\u002Fsoul\u002Fstandard`, `reve\u002Ftext-to-image`\n- Image-to-video: `higgsfield-ai\u002Fdop\u002Fstandard`, `higgsfield-ai\u002Fdop\u002Fpreview`, `bytedance\u002Fseedance\u002Fv1\u002Fpro\u002Fimage-to-video`, `kling-video\u002Fv2.1\u002Fpro\u002Fimage-to-video`\n\nMinimal Cloud payload examples:\n\n```json\n{\"prompt\":\"A serene mountain landscape at sunset\",\"aspect_ratio\":\"16:9\",\"resolution\":\"720p\"}\n```\n\n```json\n{\"image_url\":\"https:\u002F\u002Fexample.com\u002Fportrait.jpg\",\"prompt\":\"Subject turns head slightly and smiles at the camera\",\"duration\":5}\n```\n\n### Consumer CLI bootstrap\n\nOnly for the consumer CLI\u002Fdashboard surface:\n\n1. If `higgsfield` is not on `$PATH`, install it:\n   ```bash\n   curl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Fhiggsfield-ai\u002Fcli\u002Fmain\u002Finstall.sh | sh\n   ```\n2. If `higgsfield account status` fails with `Session expired` \u002F `Not authenticated`, ask the user to run `higgsfield auth login` (interactive) and wait for confirmation.\n\n\n## UX Rules\n\n1. Be concise. No raw IDs, no JSON dumps in chat. Print the media URL for generated assets, or the text summary for Virality Predictor.\n2. No internal jargon. Don't narrate \"calling higgsfield cost\", \"polling job\".\n3. Detect the user's language from the first message and reply in it. Technical args (`--aspect_ratio 16:9`) stay English.\n4. Don't batch-ask. Pick a sane default model and ask one thing at a time only if genuinely missing.\n5. Don't pre-estimate cost or optimize for cheaper models unless the user asks. Prefer the quality default first.\n6. Pass `--wait` to `generate create` so the command blocks until done and prints the result URL itself. Avoid the two-step `create` → `wait` pattern.\n\n## Discovery guardrail\n\nWhen looking for a Higgsfield feature\u002Fmodel, do not rely only on semantic search or CLI `--help`. First run an unfiltered model list, then inspect likely `job_set_type` names. If the user says a model exists but search returns no results, trust that signal and verify with the full model list before answering.\n\nWorkflows are separate from models. Discover them with `higgsfield workflow list` and inspect params with `higgsfield workflow get \u003Cworkflow_name>`.\n\nVirality Predictor is exposed as:\n\n- Customer-facing name: Virality Predictor\n- Technical `job_set_type`: `brain_activity`\n- Category\u002Foutput: text report. This is video-in\u002Ftext-out analysis, not a text\u002Fchat generation model.\n- Input: uploaded video\n- Purpose: finished-video hook, attention, retention, and virality analysis\n\nIf the user says \"analyze this video\", \"score this ad\", \"evaluate the hook\", or similar, route to `brain_activity` even though it appears under text\u002Fanalysis models. Classify by task intent and required input, not by output category alone.\n\n## Workflow — generic generation\n\nThe concrete commands in this section are consumer CLI examples. For Higgsfield Cloud API, keep the same creative\u002Fmodel choices and use the Cloud API contract selected in the surface-selection step.\n\n1. **Pick a model.** Start with the core defaults unless the brief clearly needs a specialist:\n\n   - **GPT Image 2** → default image model for high-fidelity general generation, graphic design, UI, banners, typography, and on-image text.\n   - **Seedance 2.0** → default video model for serious motion, cinematic clips, multi-shot work, image-to-video, and 4–15s production-quality output. 12s is valid.\n   - **Nano Banana 2\u002FPro** → default for character, cartoon, stylized, and reference-driven image work; use Pro for harder briefs.\n   - **Marketing Studio** → default for ads, UGC, product demos, unboxing, TV spots, presenter videos, and brand\u002Fproduct workflows.\n\n   **Image:**\n   - Brand product visual (Pinterest pin, lifestyle, hero banner, ad pack, virtual try-on) → use `higgsfield-product-photoshoot` instead. NOT this skill.\n   - Generated product concept \u002F packaging \u002F can \u002F bottle with brand name or label text → GPT Image 2.\n   - Branded ad image with avatar + product (Marketing Studio shape) → Marketing Studio Image (see Marketing Studio below)\n   - Aesthetic UGC \u002F fashion editorial \u002F lifestyle character → Soul 2.0\n   - Cinematic still frame → Soul Cinema\n   - Highly characterful creative persona (text-only, distinctive) → Soul Cast\n   - Locations \u002F environments \u002F no-people scenes → Soul Location (best in class)\n   - Logo, icon, vector-like illustration, brand mark, controlled-palette graphic → Recraft V4.1 (`recraft_v4_1`, often with `--model_type vector`)\n   - Face edit + complex scene swap → Seedream 4.5\n   - Soul Character (reference id from `higgsfield-soul-id`) → Soul 2.0 for stills, Soul Cinema for cinematic\n   - Character or cartoon-style work → Nano Banana 2; step up to Nano Banana Pro on hard cases\n   - Fast and cheap iteration → Z Image\n   - **Default for everything else → GPT Image 2.** Graphic design, UI, banners, typography, and high-fidelity general generation.\n\n   **Video:**\n   - All advertising \u002F commercial \u002F branded ad video → Marketing Studio (see Marketing Studio below)\n   - Edit existing video from sketch\u002Ftimestamp, or reframe to another aspect ratio → workflow (`draw_to_video` or `reframe`), not a model. See `references\u002Fworkflows.md`.\n   - **Default all-purpose serious video (multi-shot, consistent identity, motion-heavy, image-to-video, 4–15s requests) → Seedance 2.0.** SOTA. Do not downgrade to Seedance 1.5 just because its duration enum is easier to read; validate Seedance 2.0 first.\n   - Single-plane scene without strong dynamics, cheaper than Seedance 2.0 → Kling 3.0; if the user explicitly asks for Turbo, faster, or lower-cost Kling output → Kling 3.0 Turbo (`kling3_0_turbo`)\n   - Cheap clean shot without cuts, only when the user asks for cheaper\u002Fbudget output → Seedance 1.5 Pro\n   - Cinema-grade highest fidelity → Cinema Studio Video 3.0\n   - Cheap with strong physics, no audio needed → Minimax Hailuo\n   - Fast batch \u002F volume → Veo 3.1 Lite\n   - Bold\u002Fstylized image-to-video from a required start image → Grok Video 1.5 (`grok_video_v15`). Requires one `--start-image` or `--image`, duration 2–15s, resolution `480p` or `720p`.\n\n   **Video analysis:**\n   - Rate a finished video's hook, virality potential, attention, retention, or distraction risk → Virality Predictor (`brain_activity`). This is a video analysis model that returns a text score\u002Freport, not a generated media asset.\n\n   **3D:**\n   - Create an actual 3D mesh\u002Fmodel\u002FGLB from one or more object\u002Fproduct reference images → Multi-Image to 3D (`multi_image_to_3d`). Pass 1–4 images with repeated `--image`; use `--should_texture true` when the asset needs texture. If the user only asks for a 3D-rendered picture, use an image model instead.\n\n   **Audio:**\n   - Create non-speech sound effects, ambience, foley, impacts, or environmental audio from text → Mirelo Text to Audio (`mirelo_text_to_audio`). It requires `--prompt` and `--duration`, and returns audio. Do not pass media inputs.\n   - Create music, backing tracks, jingles, or instrumental beds from text → Sonilo Music (`sonilo_music`). It requires `--prompt` and `--duration`, and returns audio. Do not pass media inputs.\n\n   For the actual `--model` ID to pass to `higgsfield generate create`, run `higgsfield model list --json | jq` to map display names to IDs. See `references\u002Fmodel-catalog.md` for the full table.\n\n2. **Pass media inputs straight to flags.** Media flags accept a local file path **or** a UUID. CLI auto-uploads paths and auto-detects job vs upload for UUIDs. No need to pre-upload. Each model declares accepted roles (`image`, `start_image`, `end_image`, `video`, `audio`) — see `references\u002Fmedia-inputs.md`.\n3. **Validate quickly.** If unsure of params, run `higgsfield model get \u003Cjst> --json` once and pass only what's needed. Validate the preferred model before falling back to an older one. Use schema defaults otherwise. The server returns `adjustments` for non-fatal coercions (e.g. `aspect_ratio=99:99` → closest match) and a structured error for invalid declared-param values.\n4. **Submit and wait in one shot.** `higgsfield generate create \u003Cjst> [--prompt \"...\"] [media flags] [param flags] --wait`. Blocks until terminal status and prints the result on stdout. Tunables: `--wait-timeout 20m` (default 10m), `--wait-interval 5s` (default 3s). Virality Predictor does not need a prompt; pass `--video`.\n5. **Deliver.** For generated media and 3D assets, send the primary result URL plus a one-line summary (model, duration if video; GLB\u002Fasset URL for 3D). For Virality Predictor, deliver the scores, business interpretation, and the Open report link. Do not surface Virality Predictor `.glb`, `.bin`, or region-table internals in normal chat output.\n\nTo inspect or rerun later, `higgsfield generate list --json` and `higgsfield generate get \u003Cid> --json` work for retrospection. `higgsfield generate wait \u003Cid>` is still available if you ever need to rejoin a job started without `--wait`.\n\nFor workflow jobs, use `higgsfield generate workflow \u003Cworkflow_name> ... --wait`. Cost syntax is `higgsfield generate cost workflow \u003Cworkflow_name> ...`. See `references\u002Fworkflows.md`.\n\n## Media flags\n\n| Flag | Purpose | Models that accept it |\n|---|---|---|\n| `--image \u003Cpath-or-id>` | reference image | most image models, `grok_video_v15`, `multi_image_to_3d`, `seedance_2_0`, `veo3`, `marketing_studio_video` |\n| `--start-image \u003Cpath-or-id>` | first frame for image-to-video transitions | `grok_video_v15`, `kling3_0`, `kling3_0_turbo`, `kling2_6`, `veo3_1`, `seedance_2_0`, `marketing_studio_video` |\n| `--end-image \u003Cpath-or-id>` | last frame for transitions | `kling3_0`, `seedance_2_0`, `marketing_studio_video` |\n| `--video \u003Cpath-or-id>` | reference or analyzed video | `seedance_2_0`, `brain_activity` |\n| `--audio \u003Cpath-or-id>` | reference audio (lipsync, soundtrack match) | Consumer CLI path for `seedance_2_0` only. For raw\u002FAPI-style Seedance 2.0 payloads, do not send `role: \"audio\"` in `medias`; use root-level `input_audio` from `higgsfield-api-quirks`. |\n\nEach flag accepts either a local file path (auto-uploaded) or a UUID (upload id from `higgsfield upload create`, or a previous job id). Each model declares its own role set via `MEDIA_ROLES`. See `references\u002Fmedia-inputs.md` for the full table.\n\n## Common params\n\nFlags pass through to model schema. Use `higgsfield model get \u003Cjst>` to discover.\n\n```bash\nhiggsfield generate create gpt_image_2 --prompt \"neon city at dusk\" --aspect_ratio 16:9 --resolution 2k --wait\nhiggsfield generate create nano_banana_2 --prompt \"anime character concept, expressive pose\" --image .\u002Fref.png --wait\nhiggsfield generate create seedance_2_0 --prompt \"camera dollies in\" --start-image .\u002Ffirst.png --duration 12 --wait\nhiggsfield generate create grok_video_v15 --prompt \"cinematic handheld shot, neon rainy street\" --start-image .\u002Fimage.png --duration 5 --resolution 720p --wait\nhiggsfield generate create text2image_soul_v2 --prompt \"...\" --soul-id \u003Csoul_ref_id> --quality 2k --wait\nhiggsfield generate create multi_image_to_3d --image .\u002Ffront.png --image .\u002Fside.png --should_texture true --wait\nhiggsfield generate create sonilo_music --prompt \"cinematic synthwave track\" --duration 12 --wait\nhiggsfield generate create mirelo_text_to_audio --prompt \"glass breaking in a large hall\" --duration 4 --wait\nhiggsfield generate create brain_activity --video .\u002Fad.mp4 --wait\n```\n\nFor machine-readable output (chained pipelines, agent context), add `--json`. With `--wait --json` you get the final job object array. Without `--wait`, you get the job IDs. Virality Predictor stores raw analysis and render artifacts in the job params, but the default text output should stay to scores plus Open report.\n\nStdin prompt: `echo \"...\" | higgsfield generate create z_image --wait`.\n\nSoul image quality: for `text2image_soul_v2` and `soul_cinematic`, pass `--quality 1.5k` or `--quality 2k`. These are UI-facing tiers; the backend maps them to `720p`\u002F`1080p` and model-specific dimensions from the selected `--aspect_ratio`. `soul_location` has no quality selector; it uses fixed dimensions per aspect ratio.\n\n## Marketing Studio\n\nBranded image\u002Fvideo gen: avatars + products + optional setup hooks\u002Fsettings + ad-style modes. Use models `marketing_studio_video` and `marketing_studio_image`.\n\n### Concepts\n\n- **Avatar** — presenter face. Curated `preset` (browse `higgsfield marketing-studio avatars list`) or `custom` (uploaded photos via `higgsfield marketing-studio avatars create`). For UGC modes, an avatar is optional if the brief clearly mentions a person; the backend can create a Soul Character automatically. Pass an avatar when the user wants a specific presenter.\n- **Product** — brand item with title + reference images. Imported from URL (`higgsfield marketing-studio products fetch --url ...`) or created from uploaded images (`higgsfield marketing-studio products create`).\n- **Webproduct** — App Store \u002F web page version. Auto-routes when fetching App Store URLs.\n- **Hook** — reusable opening angle \u002F ad hook. Browse with `higgsfield marketing-studio hooks list`. Hook text is prepended to the user's prompt; it does not replace `--prompt`.\n- **Setting** — reusable environment \u002F scene context. Browse with `higgsfield marketing-studio settings list`.\n- **Ad reference** — reusable inspiration video that can be bound to an avatar and\u002For product. Created from an uploaded video (`--video-input \u003Cupload_id>`) or a previous generation job (`--job \u003Cjob_id>`). Browse with `higgsfield marketing-studio ad-references list`. See `references\u002Fmarketing-ad-references.md`.\n- **Brand kit** — captures a brand's identity (name, logo, hero images, colours, fonts, tone) for reuse across image generations. Created by handing in a website URL (`higgsfield marketing-studio brand-kits fetch --url https:\u002F\u002F… --wait`). See `references\u002Fmarketing-brand-kits.md`.\n- **Ad format** — presets that drives the visual structure of a generated image (`headline`, `bullet-points`, etc.). Read-only, browse with `higgsfield marketing-studio ad-formats list`. Required input for `dtc-ads generate`.\n\n### Discovery commands\n\nUse these exact list commands when the user asks what already exists:\n\n```bash\nhiggsfield marketing-studio avatars list --json\nhiggsfield marketing-studio products list --json\nhiggsfield marketing-studio hooks list --json\nhiggsfield marketing-studio settings list --json\nhiggsfield marketing-studio ad-references list --json\nhiggsfield marketing-studio brand-kits list --json\nhiggsfield marketing-studio ad-formats list --json\n```\n\n`--hook_id` and `--setting_id` are supported by `marketing_studio_video` only; do not pass them to `marketing_studio_image`.\n\n### UX rules (additional)\n\n- One question per phase. Don't ask product+avatar+mode upfront.\n- **Two ad approaches are mutually exclusive.** Either the user gives an ad reference video (reference-driven) **or** picks hook\u002Fsetting blocks (composed-from-blocks) — never both. If the user has an ad reference selected, do not offer hook\u002Fsetting; if hook\u002Fsetting are picked, do not offer to attach an ad reference.\n- **Ad reference source.** The only valid inputs are a local video file (uploaded via `higgsfield upload create ... --video`) or a prior video job. If the user provides anything else, ask for a local file.\n- **`dtc-ads` ad format is mandatory.** Always ask the user to pick from `ad-formats list`. There is no auto-default — both the CLI and server reject calls without `--format-id`.\n- **`dtc-ads` optional inputs.** Suggest avatars, products, and reference media when the brief calls for them; only attach what the user picks.\n\n### Workflow — quick ad video\n\n1. **Get product.**\n   - Existing product → `higgsfield marketing-studio products list --json`\n   - URL → `higgsfield marketing-studio products fetch --url \u003Curl> --wait` (polls until import done)\n   - Local images → `higgsfield upload create \u003Cphoto>...` then `higgsfield marketing-studio products create --title \"...\" --image \u003Cid>...`\n   Capture product id. When using `--hook_id`, strongly prefer passing `--product_ids`; hooks are designed to pivot into a product and work poorly without product context.\n2. **Pick avatar if needed.**\n   - Default: `higgsfield marketing-studio avatars list` and pick a preset matching the brand voice.\n   - Custom: `higgsfield marketing-studio avatars create --name \"...\" --image \u003Cupload_id>`.\n   For UGC modes, you may omit `--avatars` when no specific presenter is required and the brief mentions a person; the backend can synthesize a Soul Character.\n3. **Optionally pick setup items.**\n   - Hook: `higgsfield marketing-studio hooks list --json`\n   - Setting: `higgsfield marketing-studio settings list --json`\n   Pass selected IDs as `--hook_id \u003Chook_id>` and `--setting_id \u003Csetting_id>` for `marketing_studio_video` only. Do not copy the hook's prompt into `--prompt` unless the user explicitly wants to reinforce the same wording.\n4. **Pick mode if needed.** Default is `ugc`; `--mode` is not required just because `--hook_id` is present. Other current slugs: `ugc_how_to`, `ugc_unboxing`, `product_showcase`, `product_review`, `tv_spot`, `wild_card`, `ugc_virtual_try_on`, `virtual_try_on`. **Hook\u002Fsetting are valid only for `ugc`, `ugc_how_to`, `ugc_unboxing`, `product_review`, `ugc_virtual_try_on`** — do not pass `--hook_id` \u002F `--setting_id` with the other modes. See `references\u002Fmarketing-modes.md`.\n5. **Generate (one-shot).**\n   ```bash\n   PRODUCT_IDS_JSON=$(mktemp)\n   AVATARS_JSON=$(mktemp)\n   printf '[\"\u003Cproduct_id>\"]' > \"$PRODUCT_IDS_JSON\"\n   printf '[{\"id\":\"\u003Cavatar_id>\",\"type\":\"preset\"}]' > \"$AVATARS_JSON\"\n\n   higgsfield generate create marketing_studio_video \\\n     --prompt \"...\" \\\n     --avatars @\"$AVATARS_JSON\" \\\n     --product_ids @\"$PRODUCT_IDS_JSON\" \\\n     --mode ugc \\\n     --duration 15 \\\n     --resolution 720p \\\n     --aspect_ratio 9:16 \\\n     --wait\n   ```\n   Add `--hook_id \u003Chook_id>` and\u002For `--setting_id \u003Csetting_id>` when a setup hook\u002Fsetting was selected.\n   `product_ids` and `avatars` are JSON arrays; pass them via `@\u002Fpath\u002Fto\u002Ffile.json`. Do not pass a bare UUID to `--product_ids`.\n   Resolution is `480p` or `720p`. Aspect ratio is one of `auto`\u002F`21:9`\u002F`16:9`\u002F`4:3`\u002F`1:1`\u002F`3:4`\u002F`9:16`. `--generate-audio true` is supported here (unlike `seedance_2_0`). `--wait` blocks until done; bump `--wait-timeout 30m` for longer ad runs.\n6. **Deliver.** URL + one-line summary (mode, duration).\n\n### Click-to-Ad shortcut (URL-driven)\n\nWhen the user gives a product URL and wants a marketing video in one go:\n\n```bash\n# 1. Trigger fetch (returns the product id, import runs in the background)\nhiggsfield marketing-studio products fetch --url https:\u002F\u002Fshop.example.com\u002Fsneakers --wait\n\n# 2. Generate the marketing video against the same URL — backend reuses the entity\nhiggsfield generate create marketing_studio_video \\\n  --url https:\u002F\u002Fshop.example.com\u002Fsneakers \\\n  --mode ugc \\\n  --duration 15 \\\n  --aspect_ratio 9:16 \\\n  --wait\n```\n\nBackend dedupes by URL, so repeated runs reuse the existing entity instead of re-fetching.\n\n### Workflow — marketing image\n\nSame as above but use `marketing_studio_image` model:\n\n```bash\nhiggsfield generate create marketing_studio_image \\\n  --prompt \"...\" \\\n  --aspect_ratio 1:1 \\\n  --resolution 2k \\\n  --wait\n```\n\n## Virality Predictor video scoring\n\nUse Virality Predictor (`brain_activity`) when the user wants to evaluate a finished video as a business creative: hook strength, virality potential, attention, retention, or how well the content\u002Fproduct holds focus and minimizes distraction. Treat \"Virality Predictor\" as the customer-facing feature name; `brain_activity` is only the CLI\u002Fjob_set_type.\n\n```bash\nhiggsfield generate create brain_activity --video .\u002Fcreative.mp4 --wait\n```\n\nThe result is text, not a generated image\u002Fvideo. Report the overall score, peak hook second, sustain score, strongest\u002Fweakest regions, and report URL if present. Interpret it as an objective attention proxy for creative testing: higher Visual\u002FAuditory\u002FLanguage\u002FAttention scores suggest stronger stimulus and focus; lower Default Mode is better because it suggests less mind-wandering.\n\nThe CLI prints an Open report URL like `https:\u002F\u002F\u003Capp-domain>\u002Fapps\u002Fvirality-predictor?resultJobId=\u003Cjob_id>`. Send that URL for the visual report. Raw artifact URLs such as `brain_example_url`, `vertexMapBinaryUrl`, and `vertexMapUrl` are implementation details; mention them only when the user asks for raw data or implementation details.\n\nGood final shape:\n\n```text\nOverall score: 44\u002F100\nPeak hook: 49% at 1s\nSustain: 89%\nStrongest region: Visual Cortex\nRisk: Default Mode is high, which can indicate mind-wandering.\n\nOpen report: \u003Creport_url>\n```\n\n## Errors\n\n- `Missing required params: prompt` → user gave no prompt; ask for it.\n- `Missing required params: medias` on `brain_activity` \u002F Virality Predictor → pass exactly one video via `--video \u003Cpath-or-id>`.\n- `Invalid values: aspect_ratio=99:99 (allowed: ...)` → bad enum; pick from allowed.\n- `Unknown params: foo` → schema doesn't accept that flag; check `higgsfield model get \u003Cjst>`. If this happens for `hook_id` or `setting_id`, the selected model\u002Fjob_set_type does not support Marketing Studio setup items.\n- `Session expired` → `higgsfield auth login`.\n\nSee `references\u002Ftroubleshooting.md` for more.\n\n## Reference docs\n\nLoad on demand:\n\n- `references\u002Fmodel-catalog.md` — picking the right model for the task\n- `references\u002Fworkflows.md` — `draw_to_video` and `reframe` workflow generation\n- `references\u002Fprompt-engineering.md` — writing prompts that work\n- `references\u002Fmedia-inputs.md` — image\u002Fvideo\u002Faudio reference flows and Virality Predictor video analysis\n- `references\u002Ftroubleshooting.md` — common errors and fixes\n- `references\u002Fmarketing-avatars.md` — preset vs custom avatars\n- `references\u002Fmarketing-products.md` — URL fetch vs manual product create\n- `references\u002Fmarketing-setup-items.md` — hooks\u002Fsettings discovery and usage\n- `references\u002Fmarketing-ad-references.md` — ad reference videos (create\u002Flist\u002Fget)\n- `references\u002Fmarketing-brand-kits.md` — brand kits (fetch from URL, list, get)\n- `references\u002Fmarketing-dtc-ads.md` — DTC Ads Engine (`dtc-ads generate`)\n- `references\u002Fmarketing-modes.md` — every Marketing Studio mode\n","# Higgsfield Generate\n\nSubmit jobs to Higgsfield models through the right surface for the user's intent. Supports Higgsfield Cloud API jobs and the standard consumer `higgsfield` CLI\u002Fdashboard. Covers generic image\u002Fvideo\u002F3D\u002Faudio generation, Marketing Studio (branded ads, avatars, products, hooks, settings), and, secondarily, Virality Predictor video scoring.\n\n## Surface Selection - Cloud API vs Consumer CLI\n\nChoose the execution surface before any auth, install, upload, or generation step.\n\n1. If the user says \"Cloud\", \"Higgsfield Cloud\", \"API\", \"API key\", \"key id\", \"secret\", \"server-side\", \"use my env\", or \"no dashboard login\", use Higgsfield Cloud API.\n2. If the user says \"consumer\", \"standard\", \"regular Higgsfield\", \"dashboard\", \"CLI\", \"use my logged-in Higgsfield account\", or explicitly asks for the local `higgsfield` command, use the consumer CLI\u002Fdashboard.\n3. If the user has not specified and both surfaces could satisfy the request, ask one short question before proceeding: \"Use Higgsfield Cloud API or consumer Higgsfield CLI?\"\n4. Never silently switch surfaces. In particular, do not run `higgsfield auth login` for a Cloud\u002FAPI-key request, and do not fall back to procedural\u002Flocal generation when a Higgsfield generation was requested.\n\n### Cloud API execution\n\n1. Verify credentials by name only:\n   ```bash\n   hive-env-check HIGGSFIELD_API_KEY_ID\n   hive-env-check HIGGSFIELD_API_KEY_SECRET\n   ```\n2. Discover docs from the official LLM index before exploring further:\n   ```text\n   https:\u002F\u002Fdocs.higgsfield.ai\u002Fdocs\u002Fllms.txt\n   ```\n3. Use the documented Cloud REST contract:\n   - Base URL: `https:\u002F\u002Fplatform.higgsfield.ai`\n   - Submit: `POST https:\u002F\u002Fplatform.higgsfield.ai\u002F{model_id}`\n   - Status: `GET https:\u002F\u002Fplatform.higgsfield.ai\u002Frequests\u002F{request_id}\u002Fstatus`\n   - Cancel queued request: `POST https:\u002F\u002Fplatform.higgsfield.ai\u002Frequests\u002F{request_id}\u002Fcancel`\n   - Webhook: append `?hf_webhook=\u003Chttps-url>` to the model submit URL.\n   - Auth header: `Authorization: Key \u003Capi_key>:\u003Capi_secret>`\n4. Run Cloud API calls through the shared env loader and keep credentials in env expansion, not logs:\n   ```bash\n   hive-env-run -- sh -c 'curl -sS \\\n     -H \"Authorization: Key ${HIGGSFIELD_API_KEY_ID}:${HIGGSFIELD_API_KEY_SECRET}\" \\\n     -H \"Content-Type: application\u002Fjson\" \\\n     -H \"Accept: application\u002Fjson\" \\\n     -d @payload.json \\\n     https:\u002F\u002Fplatform.higgsfield.ai\u002F\u003Cmodel_id>'\n   ```\n5. Do not print, copy, log, or persist secret values. Redact token\u002Fkey\u002Fsecret fields before showing any JSON output.\n6. Prefer `curl` for direct REST checks. Python `urllib` can be blocked by Cloudflare Error 1010 (`browser_signature_banned`) before auth is evaluated. A safe no-cost auth\u002Fpath sanity check is a fake request-id status lookup; valid platform routing should return `404 {\"detail\":\"Not found\"}`, not `Missing token`, `Invalid token`, or consumer workspace credit errors:\n   ```bash\n   hive-env-run -- sh -c 'curl -sS \\\n     -H \"Accept: application\u002Fjson\" \\\n     -H \"Authorization: Key ${HIGGSFIELD_API_KEY_ID}:${HIGGSFIELD_API_KEY_SECRET}\" \\\n     https:\u002F\u002Fplatform.higgsfield.ai\u002Frequests\u002F00000000-0000-0000-0000-000000000000\u002Fstatus'\n   ```\n7. The official Python SDK is `higgsfield-client`. It accepts either `HF_KEY=\u003Capi_key>:\u003Capi_secret>` or `HF_API_KEY` plus `HF_API_SECRET`; map from the HivemindOS shared env at process launch rather than writing new secret files.\n8. Use the same model-selection guidance below, then translate prompt\u002Fmedia\u002Fmodel\u002Fduration\u002Fresolution\u002Faspect choices into the Cloud API submit\u002Fpoll\u002Fdownload lifecycle.\n9. For Seedance 2.0, Kling 3.0, or any model with known non-standard behavior, read `higgsfield-api-quirks` before building the payload. In particular, Seedance 2.0 audio must be sent as root-level `input_audio`, not as a `role: \"audio\"` item in `medias`, and aspect ratio must be the string `aspect_ratio`.\n\nCloud model examples from the official docs:\n\n- Text-to-image: `higgsfield-ai\u002Fsoul\u002Fstandard`, `reve\u002Ftext-to-image`\n- Image-to-video: `higgsfield-ai\u002Fdop\u002Fstandard`, `higgsfield-ai\u002Fdop\u002Fpreview`, `bytedance\u002Fseedance\u002Fv1\u002Fpro\u002Fimage-to-video`, `kling-video\u002Fv2.1\u002Fpro\u002Fimage-to-video`\n\nMinimal Cloud payload examples:\n\n```json\n{\"prompt\":\"A serene mountain landscape at sunset\",\"aspect_ratio\":\"16:9\",\"resolution\":\"720p\"}\n```\n\n```json\n{\"image_url\":\"https:\u002F\u002Fexample.com\u002Fportrait.jpg\",\"prompt\":\"Subject turns head slightly and smiles at the camera\",\"duration\":5}\n```\n\n### Consumer CLI bootstrap\n\nOnly for the consumer CLI\u002Fdashboard surface:\n\n1. If `higgsfield` is not on `$PATH`, install it:\n   ```bash\n   curl -fsSL https:\u002F\u002Fraw.githubusercontent.com\u002Fhiggsfield-ai\u002Fcli\u002Fmain\u002Finstall.sh | sh\n   ```\n2. If `higgsfield account status` fails with `Session expired` \u002F `Not authenticated`, ask the user to run `higgsfield auth login` (interactive) and wait for confirmation.\n\n\n## UX Rules\n\n1. Be concise. No raw IDs, no JSON dumps in chat. Print the media URL for generated assets, or the text summary for Virality Predictor.\n2. No internal jargon. Don't narrate \"calling higgsfield cost\", \"polling job\".\n3. Detect the user's language from the first message and reply in it. Technical args (`--aspect_ratio 16:9`) stay English.\n4. Don't batch-ask. Pick a sane default model and ask one thing at a time only if genuinely missing.\n5. Don't pre-estimate cost or optimize for cheaper models unless the user asks. Prefer the quality default first.\n6. Pass `--wait` to `generate create` so the command blocks until done and prints the result URL itself. Avoid the two-step `create` → `wait` pattern.\n\n## Discovery guardrail\n\nWhen looking for a Higgsfield feature\u002Fmodel, do not rely only on semantic search or CLI `--help`. First run an unfiltered model list, then inspect likely `job_set_type` names. If the user says a model exists but search returns no results, trust that signal and verify with the full model list before answering.\n\nWorkflows are separate from models. Discover them with `higgsfield workflow list` and inspect params with `higgsfield workflow get \u003Cworkflow_name>`.\n\nVirality Predictor is exposed as:\n\n- Customer-facing name: Virality Predictor\n- Technical `job_set_type`: `brain_activity`\n- Category\u002Foutput: text report. This is video-in\u002Ftext-out analysis, not a text\u002Fchat generation model.\n- Input: uploaded video\n- Purpose: finished-video hook, attention, retention, and virality analysis\n\nIf the user says \"analyze this video\", \"score this ad\", \"evaluate the hook\", or similar, route to `brain_activity` even though it appears under text\u002Fanalysis models. Classify by task intent and required input, not by output category alone.\n\n## Workflow — generic generation\n\nThe concrete commands in this section are consumer CLI examples. For Higgsfield Cloud API, keep the same creative\u002Fmodel choices and use the Cloud API contract selected in the surface-selection step.\n\n1. **Pick a model.** Start with the core defaults unless the brief clearly needs a specialist:\n\n   - **GPT Image 2** → default image model for high-fidelity general generation, graphic design, UI, banners, typography, and on-image text.\n   - **Seedance 2.0** → default video model for serious motion, cinematic clips, multi-shot work, image-to-video, and 4–15s production-quality output. 12s is valid.\n   - **Nano Banana 2\u002FPro** → default for character, cartoon, stylized, and reference-driven image work; use Pro for harder briefs.\n   - **Marketing Studio** → default for ads, UGC, product demos, unboxing, TV spots, presenter videos, and brand\u002Fproduct workflows.\n\n   **Image:**\n   - Brand product visual (Pinterest pin, lifestyle, hero banner, ad pack, virtual try-on) → use `higgsfield-product-photoshoot` instead. NOT this skill.\n   - Generated product concept \u002F packaging \u002F can \u002F bottle with brand name or label text → GPT Image 2.\n   - Branded ad image with avatar + product (Marketing Studio shape) → Marketing Studio Image (see Marketing Studio below)\n   - Aesthetic UGC \u002F fashion editorial \u002F lifestyle character → Soul 2.0\n   - Cinematic still frame → Soul Cinema\n   - Highly characterful creative persona (text-only, distinctive) → Soul Cast\n   - Locations \u002F environments \u002F no-people scenes → Soul Location (best in class)\n   - Logo, icon, vector-like illustration, brand mark, controlled-palette graphic → Recraft V4.1 (`recraft_v4_1`, often with `--model_type vector`)\n   - Face edit + complex scene swap → Seedream 4.5\n   - Soul Character (reference id from `higgsfield-soul-id`) → Soul 2.0 for stills, Soul Cinema for cinematic\n   - Character or cartoon-style work → Nano Banana 2; step up to Nano Banana Pro on hard cases\n   - Fast and cheap iteration → Z Image\n   - **Default for everything else → GPT Image 2.** Graphic design, UI, banners, typography, and high-fidelity general generation.\n\n   **Video:**\n   - All advertising \u002F commercial \u002F branded ad video → Marketing Studio (see Marketing Studio below)\n   - Edit existing video from sketch\u002Ftimestamp, or reframe to another aspect ratio → workflow (`draw_to_video` or `reframe`), not a model. See `references\u002Fworkflows.md`.\n   - **Default all-purpose serious video (multi-shot, consistent identity, motion-heavy, image-to-video, 4–15s requests) → Seedance 2.0.** SOTA. Do not downgrade to Seedance 1.5 just because its duration enum is easier to read; validate Seedance 2.0 first.\n   - Single-plane scene without strong dynamics, cheaper than Seedance 2.0 → Kling 3.0; if the user explicitly asks for Turbo, faster, or lower-cost Kling output → Kling 3.0 Turbo (`kling3_0_turbo`)\n   - Cheap clean shot without cuts, only when the user asks for cheaper\u002Fbudget output → Seedance 1.5 Pro\n   - Cinema-grade highest fidelity → Cinema Studio Video 3.0\n   - Cheap with strong physics, no audio needed → Minimax Hailuo\n   - Fast batch \u002F volume → Veo 3.1 Lite\n   - Bold\u002Fstylized image-to-video from a required start image → Grok Video 1.5 (`grok_video_v15`). Requires one `--start-image` or `--image`, duration 2–15s, resolution `480p` or `720p`.\n\n   **Video analysis:**\n   - Rate a finished video's hook, virality potential, attention, retention, or distraction risk → Virality Predictor (`brain_activity`). This is a video analysis model that returns a text score\u002Freport, not a generated media asset.\n\n   **3D:**\n   - Create an actual 3D mesh\u002Fmodel\u002FGLB from one or more object\u002Fproduct reference images → Multi-Image to 3D (`multi_image_to_3d`). Pass 1–4 images with repeated `--image`; use `--should_texture true` when the asset needs texture. If the user only asks for a 3D-rendered picture, use an image model instead.\n\n   **Audio:**\n   - Create non-speech sound effects, ambience, foley, impacts, or environmental audio from text → Mirelo Text to Audio (`mirelo_text_to_audio`). It requires `--prompt` and `--duration`, and returns audio. Do not pass media inputs.\n   - Create music, backing tracks, jingles, or instrumental beds from text → Sonilo Music (`sonilo_music`). It requires `--prompt` and `--duration`, and returns audio. Do not pass media inputs.\n\n   For the actual `--model` ID to pass to `higgsfield generate create`, run `higgsfield model list --json | jq` to map display names to IDs. See `references\u002Fmodel-catalog.md` for the full table.\n\n2. **Pass media inputs straight to flags.** Media flags accept a local file path **or** a UUID. CLI auto-uploads paths and auto-detects job vs upload for UUIDs. No need to pre-upload. Each model declares accepted roles (`image`, `start_image`, `end_image`, `video`, `audio`) — see `references\u002Fmedia-inputs.md`.\n3. **Validate quickly.** If unsure of params, run `higgsfield model get \u003Cjst> --json` once and pass only what's needed. Validate the preferred model before falling back to an older one. Use schema defaults otherwise. The server returns `adjustments` for non-fatal coercions (e.g. `aspect_ratio=99:99` → closest match) and a structured error for invalid declared-param values.\n4. **Submit and wait in one shot.** `higgsfield generate create \u003Cjst> [--prompt \"...\"] [media flags] [param flags] --wait`. Blocks until terminal status and prints the result on stdout. Tunables: `--wait-timeout 20m` (default 10m), `--wait-interval 5s` (default 3s). Virality Predictor does not need a prompt; pass `--video`.\n5. **Deliver.** For generated media and 3D assets, send the primary result URL plus a one-line summary (model, duration if video; GLB\u002Fasset URL for 3D). For Virality Predictor, deliver the scores, business interpretation, and the Open report link. Do not surface Virality Predictor `.glb`, `.bin`, or region-table internals in normal chat output.\n\nTo inspect or rerun later, `higgsfield generate list --json` and `higgsfield generate get \u003Cid> --json` work for retrospection. `higgsfield generate wait \u003Cid>` is still available if you ever need to rejoin a job started without `--wait`.\n\nFor workflow jobs, use `higgsfield generate workflow \u003Cworkflow_name> ... --wait`. Cost syntax is `higgsfield generate cost workflow \u003Cworkflow_name> ...`. See `references\u002Fworkflows.md`.\n\n## Media flags\n\n| Flag | Purpose | Models that accept it |\n|---|---|---|\n| `--image \u003Cpath-or-id>` | reference image | most image models, `grok_video_v15`, `multi_image_to_3d`, `seedance_2_0`, `veo3`, `marketing_studio_video` |\n| `--start-image \u003Cpath-or-id>` | first frame for image-to-video transitions | `grok_video_v15`, `kling3_0`, `kling3_0_turbo`, `kling2_6`, `veo3_1`, `seedance_2_0`, `marketing_studio_video` |\n| `--end-image \u003Cpath-or-id>` | last frame for transitions | `kling3_0`, `seedance_2_0`, `marketing_studio_video` |\n| `--video \u003Cpath-or-id>` | reference or analyzed video | `seedance_2_0`, `brain_activity` |\n| `--audio \u003Cpath-or-id>` | reference audio (lipsync, soundtrack match) | Consumer CLI path for `seedance_2_0` only. For raw\u002FAPI-style Seedance 2.0 payloads, do not send `role: \"audio\"` in `medias`; use root-level `input_audio` from `higgsfield-api-quirks`. |\n\nEach flag accepts either a local file path (auto-uploaded) or a UUID (upload id from `higgsfield upload create`, or a previous job id). Each model declares its own role set via `MEDIA_ROLES`. See `references\u002Fmedia-inputs.md` for the full table.\n\n## Common params\n\nFlags pass through to model schema. Use `higgsfield model get \u003Cjst>` to discover.\n\n```bash\nhiggsfield generate create gpt_image_2 --prompt \"neon city at dusk\" --aspect_ratio 16:9 --resolution 2k --wait\nhiggsfield generate create nano_banana_2 --prompt \"anime character concept, expressive pose\" --image .\u002Fref.png --wait\nhiggsfield generate create seedance_2_0 --prompt \"camera dollies in\" --start-image .\u002Ffirst.png --duration 12 --wait\nhiggsfield generate create grok_video_v15 --prompt \"cinematic handheld shot, neon rainy street\" --start-image .\u002Fimage.png --duration 5 --resolution 720p --wait\nhiggsfield generate create text2image_soul_v2 --prompt \"...\" --soul-id \u003Csoul_ref_id> --quality 2k --wait\nhiggsfield generate create multi_image_to_3d --image .\u002Ffront.png --image .\u002Fside.png --should_texture true --wait\nhiggsfield generate create sonilo_music --prompt \"cinematic synthwave track\" --duration 12 --wait\nhiggsfield generate create mirelo_text_to_audio --prompt \"glass breaking in a large hall\" --duration 4 --wait\nhiggsfield generate create brain_activity --video .\u002Fad.mp4 --wait\n```\n\nFor machine-readable output (chained pipelines, agent context), add `--json`. With `--wait --json` you get the final job object array. Without `--wait`, you get the job IDs. Virality Predictor stores raw analysis and render artifacts in the job params, but the default text output should stay to scores plus Open report.\n\nStdin prompt: `echo \"...\" | higgsfield generate create z_image --wait`.\n\nSoul image quality: for `text2image_soul_v2` and `soul_cinematic`, pass `--quality 1.5k` or `--quality 2k`. These are UI-facing tiers; the backend maps them to `720p`\u002F`1080p` and model-specific dimensions from the selected `--aspect_ratio`. `soul_location` has no quality selector; it uses fixed dimensions per aspect ratio.\n\n## Marketing Studio\n\nBranded image\u002Fvideo gen: avatars + products + optional setup hooks\u002Fsettings + ad-style modes. Use models `marketing_studio_video` and `marketing_studio_image`.\n\n### Concepts\n\n- **Avatar** — presenter face. Curated `preset` (browse `higgsfield marketing-studio avatars list`) or `custom` (uploaded photos via `higgsfield marketing-studio avatars create`). For UGC modes, an avatar is optional if the brief clearly mentions a person; the backend can create a Soul Character automatically. Pass an avatar when the user wants a specific presenter.\n- **Product** — brand item with title + reference images. Imported from URL (`higgsfield marketing-studio products fetch --url ...`) or created from uploaded images (`higgsfield marketing-studio products create`).\n- **Webproduct** — App Store \u002F web page version. Auto-routes when fetching App Store URLs.\n- **Hook** — reusable opening angle \u002F ad hook. Browse with `higgsfield marketing-studio hooks list`. Hook text is prepended to the user's prompt; it does not replace `--prompt`.\n- **Setting** — reusable environment \u002F scene context. Browse with `higgsfield marketing-studio settings list`.\n- **Ad reference** — reusable inspiration video that can be bound to an avatar and\u002For product. Created from an uploaded video (`--video-input \u003Cupload_id>`) or a previous generation job (`--job \u003Cjob_id>`). Browse with `higgsfield marketing-studio ad-references list`. See `references\u002Fmarketing-ad-references.md`.\n- **Brand kit** — captures a brand's identity (name, logo, hero images, colours, fonts, tone) for reuse across image generations. Created by handing in a website URL (`higgsfield marketing-studio brand-kits fetch --url https:\u002F\u002F… --wait`). See `references\u002Fmarketing-brand-kits.md`.\n- **Ad format** — presets that drives the visual structure of a generated image (`headline`, `bullet-points`, etc.). Read-only, browse with `higgsfield marketing-studio ad-formats list`. Required input for `dtc-ads generate`.\n\n### Discovery commands\n\nUse these exact list commands when the user asks what already exists:\n\n```bash\nhiggsfield marketing-studio avatars list --json\nhiggsfield marketing-studio products list --json\nhiggsfield marketing-studio hooks list --json\nhiggsfield marketing-studio settings list --json\nhiggsfield marketing-studio ad-references list --json\nhiggsfield marketing-studio brand-kits list --json\nhiggsfield marketing-studio ad-formats list --json\n```\n\n`--hook_id` and `--setting_id` are supported by `marketing_studio_video` only; do not pass them to `marketing_studio_image`.\n\n### UX rules (additional)\n\n- One question per phase. Don't ask product+avatar+mode upfront.\n- **Two ad approaches are mutually exclusive.** Either the user gives an ad reference video (reference-driven) **or** picks hook\u002Fsetting blocks (composed-from-blocks) — never both. If the user has an ad reference selected, do not offer hook\u002Fsetting; if hook\u002Fsetting are picked, do not offer to attach an ad reference.\n- **Ad reference source.** The only valid inputs are a local video file (uploaded via `higgsfield upload create ... --video`) or a prior video job. If the user provides anything else, ask for a local file.\n- **`dtc-ads` ad format is mandatory.** Always ask the user to pick from `ad-formats list`. There is no auto-default — both the CLI and server reject calls without `--format-id`.\n- **`dtc-ads` optional inputs.** Suggest avatars, products, and reference media when the brief calls for them; only attach what the user picks.\n\n### Workflow — quick ad video\n\n1. **Get product.**\n   - Existing product → `higgsfield marketing-studio products list --json`\n   - URL → `higgsfield marketing-studio products fetch --url \u003Curl> --wait` (polls until import done)\n   - Local images → `higgsfield upload create \u003Cphoto>...` then `higgsfield marketing-studio products create --title \"...\" --image \u003Cid>...`\n   Capture product id. When using `--hook_id`, strongly prefer passing `--product_ids`; hooks are designed to pivot into a product and work poorly without product context.\n2. **Pick avatar if needed.**\n   - Default: `higgsfield marketing-studio avatars list` and pick a preset matching the brand voice.\n   - Custom: `higgsfield marketing-studio avatars create --name \"...\" --image \u003Cupload_id>`.\n   For UGC modes, you may omit `--avatars` when no specific presenter is required and the brief mentions a person; the backend can synthesize a Soul Character.\n3. **Optionally pick setup items.**\n   - Hook: `higgsfield marketing-studio hooks list --json`\n   - Setting: `higgsfield marketing-studio settings list --json`\n   Pass selected IDs as `--hook_id \u003Chook_id>` and `--setting_id \u003Csetting_id>` for `marketing_studio_video` only. Do not copy the hook's prompt into `--prompt` unless the user explicitly wants to reinforce the same wording.\n4. **Pick mode if needed.** Default is `ugc`; `--mode` is not required just because `--hook_id` is present. Other current slugs: `ugc_how_to`, `ugc_unboxing`, `product_showcase`, `product_review`, `tv_spot`, `wild_card`, `ugc_virtual_try_on`, `virtual_try_on`. **Hook\u002Fsetting are valid only for `ugc`, `ugc_how_to`, `ugc_unboxing`, `product_review`, `ugc_virtual_try_on`** — do not pass `--hook_id` \u002F `--setting_id` with the other modes. See `references\u002Fmarketing-modes.md`.\n5. **Generate (one-shot).**\n   ```bash\n   PRODUCT_IDS_JSON=$(mktemp)\n   AVATARS_JSON=$(mktemp)\n   printf '[\"\u003Cproduct_id>\"]' > \"$PRODUCT_IDS_JSON\"\n   printf '[{\"id\":\"\u003Cavatar_id>\",\"type\":\"preset\"}]' > \"$AVATARS_JSON\"\n\n   higgsfield generate create marketing_studio_video \\\n     --prompt \"...\" \\\n     --avatars @\"$AVATARS_JSON\" \\\n     --product_ids @\"$PRODUCT_IDS_JSON\" \\\n     --mode ugc \\\n     --duration 15 \\\n     --resolution 720p \\\n     --aspect_ratio 9:16 \\\n     --wait\n   ```\n   Add `--hook_id \u003Chook_id>` and\u002For `--setting_id \u003Csetting_id>` when a setup hook\u002Fsetting was selected.\n   `product_ids` and `avatars` are JSON arrays; pass them via `@\u002Fpath\u002Fto\u002Ffile.json`. Do not pass a bare UUID to `--product_ids`.\n   Resolution is `480p` or `720p`. Aspect ratio is one of `auto`\u002F`21:9`\u002F`16:9`\u002F`4:3`\u002F`1:1`\u002F`3:4`\u002F`9:16`. `--generate-audio true` is supported here (unlike `seedance_2_0`). `--wait` blocks until done; bump `--wait-timeout 30m` for longer ad runs.\n6. **Deliver.** URL + one-line summary (mode, duration).\n\n### Click-to-Ad shortcut (URL-driven)\n\nWhen the user gives a product URL and wants a marketing video in one go:\n\n```bash\n# 1. Trigger fetch (returns the product id, import runs in the background)\nhiggsfield marketing-studio products fetch --url https:\u002F\u002Fshop.example.com\u002Fsneakers --wait\n\n# 2. Generate the marketing video against the same URL — backend reuses the entity\nhiggsfield generate create marketing_studio_video \\\n  --url https:\u002F\u002Fshop.example.com\u002Fsneakers \\\n  --mode ugc \\\n  --duration 15 \\\n  --aspect_ratio 9:16 \\\n  --wait\n```\n\nBackend dedupes by URL, so repeated runs reuse the existing entity instead of re-fetching.\n\n### Workflow — marketing image\n\nSame as above but use `marketing_studio_image` model:\n\n```bash\nhiggsfield generate create marketing_studio_image \\\n  --prompt \"...\" \\\n  --aspect_ratio 1:1 \\\n  --resolution 2k \\\n  --wait\n```\n\n## Virality Predictor video scoring\n\nUse Virality Predictor (`brain_activity`) when the user wants to evaluate a finished video as a business creative: hook strength, virality potential, attention, retention, or how well the content\u002Fproduct holds focus and minimizes distraction. Treat \"Virality Predictor\" as the customer-facing feature name; `brain_activity` is only the CLI\u002Fjob_set_type.\n\n```bash\nhiggsfield generate create brain_activity --video .\u002Fcreative.mp4 --wait\n```\n\nThe result is text, not a generated image\u002Fvideo. Report the overall score, peak hook second, sustain score, strongest\u002Fweakest regions, and report URL if present. Interpret it as an objective attention proxy for creative testing: higher Visual\u002FAuditory\u002FLanguage\u002FAttention scores suggest stronger stimulus and focus; lower Default Mode is better because it suggests less mind-wandering.\n\nThe CLI prints an Open report URL like `https:\u002F\u002F\u003Capp-domain>\u002Fapps\u002Fvirality-predictor?resultJobId=\u003Cjob_id>`. Send that URL for the visual report. Raw artifact URLs such as `brain_example_url`, `vertexMapBinaryUrl`, and `vertexMapUrl` are implementation details; mention them only when the user asks for raw data or implementation details.\n\nGood final shape:\n\n```text\nOverall score: 44\u002F100\nPeak hook: 49% at 1s\nSustain: 89%\nStrongest region: Visual Cortex\nRisk: Default Mode is high, which can indicate mind-wandering.\n\nOpen report: \u003Creport_url>\n```\n\n## Errors\n\n- `Missing required params: prompt` → user gave no prompt; ask for it.\n- `Missing required params: medias` on `brain_activity` \u002F Virality Predictor → pass exactly one video via `--video \u003Cpath-or-id>`.\n- `Invalid values: aspect_ratio=99:99 (allowed: ...)` → bad enum; pick from allowed.\n- `Unknown params: foo` → schema doesn't accept that flag; check `higgsfield model get \u003Cjst>`. If this happens for `hook_id` or `setting_id`, the selected model\u002Fjob_set_type does not support Marketing Studio setup items.\n- `Session expired` → `higgsfield auth login`.\n\nSee `references\u002Ftroubleshooting.md` for more.\n\n## Reference docs\n\nLoad on demand:\n\n- `references\u002Fmodel-catalog.md` — picking the right model for the task\n- `references\u002Fworkflows.md` — `draw_to_video` and `reframe` workflow generation\n- `references\u002Fprompt-engineering.md` — writing prompts that work\n- `references\u002Fmedia-inputs.md` — image\u002Fvideo\u002Faudio reference flows and Virality Predictor video analysis\n- `references\u002Ftroubleshooting.md` — common errors and fixes\n- `references\u002Fmarketing-avatars.md` — preset vs custom avatars\n- `references\u002Fmarketing-products.md` — URL fetch vs manual product create\n- `references\u002Fmarketing-setup-items.md` — hooks\u002Fsettings discovery and usage\n- `references\u002Fmarketing-ad-references.md` — ad reference videos (create\u002Flist\u002Fget)\n- `references\u002Fmarketing-brand-kits.md` — brand kits (fetch from URL, list, get)\n- `references\u002Fmarketing-dtc-ads.md` — DTC Ads Engine (`dtc-ads generate`)\n- `references\u002Fmarketing-modes.md` — every Marketing Studio mode\n",{"name":5,"version":24,"description":25,"allowed-tools":26,"argument-hint":27},"0.3.0","Generate images\u002Fvideos\u002F3D assets\u002Faudio via Higgsfield AI. Defaults:\nGPT Image 2 for image\u002Fdesign\u002Ftext, Seedance 2.0 for\nvideo, Nano Banana 2\u002FPro for character\u002Freference images,\nMarketing Studio for ads, Sonilo\u002FMirelo for audio, plus\nSoul models and Kling 3.0.\nUse when: \"generate an image\", \"make a video\", \"animate\nthis photo\", \"image-to-video\", \"edit\u002Fstylize\u002Fremix this\nimage\", \"produce a clip\", \"reframe this video\", \"edit\nthis video from a sketch\", \"create a 3D model\", \"make a\nGLB\u002Fmesh\", \"create a sound effect\", \"make music\",\n\"text-to-audio\", \"create an ad\", \"make a UGC video\",\n\"product demo\", \"unboxing\", \"brand video\", \"presenter\nvideo\", \"import product from URL\", \"create avatar for ad\",\nor \"analyze video virality\". Supports image-to-image,\nimage-to-video, image-to-3D (`multi_image_to_3d`),\ntext-to-audio (`mirelo_text_to_audio`), text-to-music\n(`sonilo_music`), workflow generation (`draw_to_video`,\n`reframe`), references, job\u002Fupload IDs, Marketing Studio,\nand Virality Predictor (`brain_activity`).\nSupports two execution surfaces: Higgsfield Cloud API\nwith HIGGSFIELD_API_KEY_ID\u002FHIGGSFIELD_API_KEY_SECRET,\nand the standard consumer Higgsfield CLI\u002Fdashboard.\nChain with higgsfield-soul-id for face\u002Fidentity consistency.\nNOT for: Soul Character training (use higgsfield-soul-id),\nproduct photoshoots, marketplace listing cards,\ntext\u002Fchat\u002FTTS tasks.\n","Bash","[prompt-or-analysis-request] [--model \u003Cname>] [--image|--video \u003Cpath-or-id>]","main","f270c4ae4f9d09c4e5da9cfa8adab7e89ae1f53e","packaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate",4,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos","https:\u002F\u002Fgithub.com\u002FLiamVisionary","2026-11-10 21:48:41",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":47},[37,40,43],{"tool":38,"command":39},"npx","npx skills add LiamVisionary\u002Fhivemindos --skill higgsfield-generate",{"tool":41,"command":42},"bunx","bunx skills add LiamVisionary\u002Fhivemindos --skill higgsfield-generate",{"tool":44,"command":45},"pnpm","pnpm dlx skills add LiamVisionary\u002Fhivemindos --skill higgsfield-generate","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002FSKILL.md",true,[49,55,59,63,67,71,75,79,83,87,91,95,99,103],{"path":50,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":53,"sourceHtmlUrl":54},".hivemind-skill-source.json","REFERENCE","参考文档",965,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002F.hivemind-skill-source.json",{"path":56,"resourceType":56,"resourceTypeDesc":57,"sizeBytes":58,"sourceHtmlUrl":15},"LICENSE","许可证",1070,{"path":60,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":61,"sourceHtmlUrl":62},"references\u002Fmarketing-ad-references.md",4314,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fmarketing-ad-references.md",{"path":64,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":65,"sourceHtmlUrl":66},"references\u002Fmarketing-avatars.md",1395,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fmarketing-avatars.md",{"path":68,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":69,"sourceHtmlUrl":70},"references\u002Fmarketing-brand-kits.md",2256,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fmarketing-brand-kits.md",{"path":72,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":73,"sourceHtmlUrl":74},"references\u002Fmarketing-dtc-ads.md",4698,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fmarketing-dtc-ads.md",{"path":76,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":77,"sourceHtmlUrl":78},"references\u002Fmarketing-modes.md",3330,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fmarketing-modes.md",{"path":80,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":81,"sourceHtmlUrl":82},"references\u002Fmarketing-products.md",1505,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fmarketing-products.md",{"path":84,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":85,"sourceHtmlUrl":86},"references\u002Fmarketing-setup-items.md",2640,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fmarketing-setup-items.md",{"path":88,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":89,"sourceHtmlUrl":90},"references\u002Fmedia-inputs.md",6599,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fmedia-inputs.md",{"path":92,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":93,"sourceHtmlUrl":94},"references\u002Fmodel-catalog.md",18141,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fmodel-catalog.md",{"path":96,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":97,"sourceHtmlUrl":98},"references\u002Fprompt-engineering.md",1554,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fprompt-engineering.md",{"path":100,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":101,"sourceHtmlUrl":102},"references\u002Ftroubleshooting.md",1593,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Ftroubleshooting.md",{"path":104,"resourceType":51,"resourceTypeDesc":52,"sizeBytes":105,"sourceHtmlUrl":106},"references\u002Fworkflows.md",3059,"https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fpackaged-skills\u002Foptional\u002Fmedia\u002Fhiggsfield\u002Fhiggsfield-generate\u002Freferences\u002Fworkflows.md",[108,115,122,129,136,143],{"skillNo":109,"name":110,"nameZh":6,"description":111,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":19},"01M734AA85KQEXDQXQJ3T7769M","tailscale-generation-skill-authoring","This skill should be used when the user asks to \"create a Tailscale generation skill\", \"add a local generation skill\", \"document a generation endpoint\", \"add an image generation provider\", \"add a TTS provider\", \"add video generation\", \"add 3D generation\", or needs guidance on where HivemindOS Tailscale or local generation providers belong in the shared skills shelf.","https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fskills\u002Ftailscale-generation-skill-authoring\u002FSKILL.md",[],[],{"skillNo":116,"name":117,"nameZh":6,"description":118,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":119,"categories":120,"keywords":121,"publishedAt":19},"01M734A7VD69PRA0C0EV5EFY9T","localtts","This skill should be used when the user asks to \"generate local TTS\", \"use local text to speech\", \"use remote TTS over Tailscale\", \"make speech audio\", \"add a TTS provider\", \"use Qwen3-TTS\", or needs TTS endpoint setup through local or HivemindOS Tailscale routes.","https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fskills\u002Flocaltts\u002FSKILL.md",[],[],{"skillNo":123,"name":124,"nameZh":6,"description":125,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":126,"categories":127,"keywords":128,"publishedAt":19},"01M734A4K2Z0F74QZP7V7WBFRY","karpathy-guidelines","Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.","https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fskills\u002Fkarpathy-guidelines\u002FSKILL.md",[],[],{"skillNo":130,"name":131,"nameZh":6,"description":132,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":133,"categories":134,"keywords":135,"publishedAt":19},"01M734A1HMVYVJX3YJGNQ6KF6C","comfyui-image-generation","This skill should be used when the user asks to \"generate images\", \"generate images with ComfyUI\", \"use local ComfyUI\", \"use remote ComfyUI\", \"use Z-Image\", \"add an image generation provider\", or needs image-generation endpoint setup through local or HivemindOS Tailscale routes.","https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fskills\u002Fcomfyui-image-generation\u002FSKILL.md",[],[],{"skillNo":137,"name":138,"nameZh":6,"description":139,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":140,"categories":141,"keywords":142,"publishedAt":19},"01M7349Y684Q676DNNHZPSC8W1","wrangler","Cloudflare Workers CLI for deploying, developing, and managing Workers, KV, R2, D1, Vectorize, Hyperdrive, Workers AI, Containers, Queues, Workflows, Pipelines, and Secrets Store. Load before running wrangler commands to ensure correct syntax and best practices. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.","https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fskills\u002Fcloudflare\u002Fwrangler\u002FSKILL.md",[],[],{"skillNo":144,"name":145,"nameZh":6,"description":146,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":147,"categories":148,"keywords":149,"publishedAt":19},"01M7349V3WCBF835QP74SA9Q8C","workers-best-practices","Reviews and authors Cloudflare Workers code against production best practices. Load when writing new Workers, reviewing Worker code, configuring wrangler.jsonc, or checking for common Workers anti-patterns (streaming, floating promises, global state, secrets, bindings, observability). Biases towards retrieval from Cloudflare docs over pre-trained knowledge.","https:\u002F\u002Fgithub.com\u002FLiamVisionary\u002Fhivemindos\u002Fblob\u002Fmain\u002Fskills\u002Fcloudflare\u002Fworkers-best-practices\u002FSKILL.md",[],[]]