Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/OSideMedia/higgsfield-ai-prompt-skillnpx agentmods add skills/osidemedia/higgsfield-ai-prompt-skill/higgsfield-seedance-2-5Wrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/osidemedia/higgsfield-ai-prompt-skill/higgsfield-seedance-2-5)<a href="https://agentmods.dev/skills/osidemedia/higgsfield-ai-prompt-skill/higgsfield-seedance-2-5"><img src="https://agentmods.dev/badge/skills/osidemedia/higgsfield-ai-prompt-skill/higgsfield-seedance-2-5/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/osidemedia/higgsfield-ai-prompt-skill/higgsfield-seedance-2-5"><img src="https://agentmods.dev/badge/skills/osidemedia/higgsfield-ai-prompt-skill/higgsfield-seedance-2-5.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00191 | $0.09975 |
| Opus 5 | $0.00096 | $0.04987 |
| Sonnet 5 | $0.00038 | $0.01995 |
| Haiku 4.5 | $0.00019 | $0.00997 |
Grade A, and why
higgsfield-seedance-2-5 scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 709 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Higgsfield Seedance 2.5 Director
Seedance 2.5 is a different dialect from Seedance 2.0, not a version bump you can prompt through by habit. 2.0 is a reference-driven shot generator with start/end frames and a 4K lane. 2.5 is an omni-reference production model: up to 50 reference materials, 30-second native runtime, and three non-generation modes — it can edit a video you already have, and extend one forward or backward from its boundary frame.
The prompt grammar changes with it. Reference roles are declared in prose (@Image 1 defines …), audio and text get bracket syntax, long videos are staged with explicit end
states, and first/last frames are announced inside the prompt rather than selected as
a mode.
Model split — read this before writing anything. 2.5 caps at 720p and has no
start_image/end_imagemedia role, nogenrehint, and no 4K lane. If the job needs 4K, a genre hint, or platform-level start/end frame pinning, it is a Seedance 2.0 job —../higgsfield-seedance/SKILL.md. See § Choosing 2.0 vs 2.5.
QUICK FACTS
Generated-checked block (scripts/build_index.py verifies anchors). Routing aids — read the linked sections for the rules themselves.
- Four modes, picked before writing:
t2v·omni_reference·video_edit·video_extension; the mode changes what the prompt is → - Higgsfield surface: 480p/720p only, duration 4–30s, no start/end-frame role, no genre hint,
extension_moderequired for (and only for)video_extension→ video_editignoresdurationandaspect_ratioand bills by the source video's length;video_extensioninherits the source's aspect ratio →- Every reference material gets an explicit role and an exclusion — "what to use" plus "what not to use"; never let the model infer the mapping →
- Each material also declares a fidelity grade — full-preserve / partial-preserve / attribute-transfer (name the target) / loose-guide; beat lines name characters (name + one visible marker), never handles →
- Material budget: 30 images / 10 videos ≤30s total / 10 audio ≤30s total, 50 materials max; stability ranges are 1–8 subjects (images), 1–5 subjects at 5–10s (video/audio) →
- Multi-reference is a 5-step workflow — map → group → profile → select-by-scene, one line per subject;
@Images 1 through 4 define four charactersis the canonical failure → - Long videos are staged, not paragraphed: one primary change per stage + an explicit end state; timestamps allocate a budget, they are not frame-accurate edit points →
- Staging fixes too many EVENTS; two incompatible JOBS in one generation (physics + performance) is a separate cut — split into two prompts and stitch →
- Bracket syntax:
()music ·<>SFX ·{}dialogue ·【】subtitles; non-Chinese dialogue needs a language line before the line → - First/last frames and multi-keyframes are declared in the prompt (
@Image 1 is the first frame) — aspect ratio locks to the first image; never merge the two anchors into one sentence → - Editing needs a sole editing master + edit scope + Timeline Inheritance; extension needs the boundary frame aligned before any new content:
MODE-PLAYBOOKS.md - Storyboard grids, coarse-vs-fine blockouts, one-click video, seamless transitions:
MODE-PLAYBOOKS.md - AI-VFX production pipeline — model-per-asset-class routing, the size-ref frame, location batching, the
omni_referencev2v lane (source ≥4s, duration = source), the four-batch rule, the slop catalog:VFX-PIPELINE.md - Emotion needs 2–4 observable cues, not adjectives; niche camera terms get translated into a visible result →
- The real-person formula is 7 slots — and slot 1 is role, never age: the age-blind engine rule outranks the source guide's
[Age/Race]label → - Hard limits that must not be over-promised (frame accuracy, locked parameters, pixel-identical transitions) →
- Dreamina-product features that are not on the Higgsfield surface — Ultra Long Video 180s, mark-based editing, Clay Renderer →
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 709 lines · 191 tokens per session scan A 9d61bac33324
higgsfield-seedance-2-5 is a skill published in the GitHub repository OSideMedia/higgsfield-ai-prompt-skill (531 stars, last pushed 20d ago), licensed MIT. It adds 191 tokens to every session and 9,975 once invoked, about $0.0010 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
vibe-creating-prompt
Judges whether a user's input suits the Vibe Creating style of video-prompt writing, and when it does, distills single-scene prompts, multi-shot descriptions, emotional imagery, or mixed input into prompts that are easier for a video model to generate from — while preserving any user-specified dialogue, voiceover…
seedance-cinematic-film
Write photorealistic live-action cinematic Seedance 2.0 prompts for Higgsfield, built on five grounding pillars that stop AI drift and floaty motion. Use for "cinematic film prompt", "shot like a movie", realistic body movement, grounded motion, restrained emotional close-ups, driving scenes, fight choreography…
director-cinematography
Virtual director and cinematographer for AI-generated short videos using Higgsfield. Takes a script, scene, or creative idea and produces a shot-by-shot visual direction plan with ready-to-paste Higgsfield prompts — camera presets, lenses, lighting, color grading, and motion directives for every shot. Channels 11…
seedance-prompt-builder
Write production-ready Seedance 2.0 prompts (image, video, and motion/edit) for Higgsfield. Use this skill whenever the user wants a single Seedance prompt rather than a full shotlist — e.g. "make a Seedance prompt for X", "write me a video prompt", "restyle this clip", "turn my photo into a character sheet"…
down-skilling
Distill Opus-level reasoning into optimized instructions for Haiku 4.5 (and Sonnet). Generates explicit, procedural prompts with n-shot examples that maximize smaller model performance on a given task. Use when user says "down-skill", "distill for Haiku", "optimize for Haiku", "make this work on Haiku", "generate…
reviewing-ai-papers
Analyzes an AI/ML publication — paper, preprint, article, technical blog post — and extracts what an enterprise AI engineer should do about it. Use when someone supplies a URL or document on RAG, embeddings, fine-tuning, prompt engineering, agents, or LLM deployment and asks "review this paper", "what do you make of…