higgsfield
OSideMedia/higgsfield-ai-prompt-skill
writes your Higgsfield prompts, picks model and camera , which improves: fewer failed generations
Details
- judge’s score
- 8/10
- place
- outside the packs
- publisher
- OSideMedia
- status
- new
- pinned commit
- 7075497 on GitHub (new tab)
- skill’s last commit
- Sep 27, 2026
- repo stars
- 767
- main type
- Video
Judgement
Very clear triggering and precise Higgsfield vocabulary, with dated snapshots and verifiable rules. Very broad scope (routing to about fifty sub-skills), excessive length, and a style imposed by default.
reservations
- SKILL.md truncated at 24000 characters out of 36348: the end (routing, templates) was not read.
- Cited sub-skills and files are absent (skills/shared/negative-constraints.md, specs/model-specs.yaml, templates/, higgsfield-* sub-skills): the references cannot be verified.
- The Fast Path imposes a Cinematic style by default and forbids questions, which decides the aesthetics in place of the user.
- Very broad scope and volatile catalogue (snapshot of 2026-09-26, many models marked "unverified"), with many meta-rules that make it heavier to read.
- Triggering
- 2/2
- Added value
- 2/2
- Precision
- 2/2
- Scope
- 1/2
- Quality
- 1/2
Alerts
1 alert, by category
Network
- high risk
The skill sends prompts, scripts and reference files (faces, characters) to the third-party service Higgsfield via CLI or MCP, with a connection through "higgsfield auth login". This data flow is not disclosed to the user.
Original description
Use this skill whenever the user asks anything about Higgsfield AI — writing or refining video/image prompts, choosing a model (Kling, Veo, Wan, Seedance, Minimax Hailuo, DoP, Soul, Nano Banana, Seedream, Flux, GPT Image, etc.), camera controls, named motion presets, Soul ID character consistency, C…
added on Oct 19, 2026 · license MIT