Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/slavasexton/comfyui-agent-kit/kreanpx skills add SlavaSexton/ComfyUI-Agent-Kit --skill kreagit clone --depth 1 https://github.com/SlavaSexton/ComfyUI-Agent-KitWrote 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/slavasexton/comfyui-agent-kit/krea)<a href="https://agentmods.dev/skills/slavasexton/comfyui-agent-kit/krea"><img src="https://agentmods.dev/badge/skills/slavasexton/comfyui-agent-kit/krea.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00104 | $0.03091 |
| Opus 5 | $0.00052 | $0.01545 |
| Sonnet 5 | $0.00021 | $0.00618 |
| Haiku 4.5 | $0.00010 | $0.00309 |
Grade A, and why
krea 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 6d 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 — 160 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Krea
Krea ships two completely separate things under one brand, and picking the wrong one wastes the job.
The fork, decide it first.
- Hosted API. Nodes
Krea2ImageNodeandKrea2StyleReferenceNode, categorypartner/image/Krea, billed per image, capped at 1K output, and the only path that reads a Krea moodboard from the website. - Open weights. Krea 2 Raw and Turbo run locally on core nodes, free, any resolution, no moodboard.
Who owns what, so you open one file, not three. THIS file owns the fork, the hosted graph, the FLUX.1 Krea
Dev graph and the ecosystem map. reference.md next to it owns full node I/O, the price matrix, weight files
with measured sizes, and the custom-node packs. The kit's MODELS.md already owns the local Krea 2 graph in
depth (Raw vs Turbo settings, the qwen3vl_4b encoder and its krea2 CLIP type, the Ostris edit method, the
style-reference LoRA on core nodes, the nine official style LoRAs, the abliterated-encoder swap, licence).
Do not restate it here, route to it.
Which Krea for which job
| Job | Path | Why this one |
|---|---|---|
| A look you already built on krea.ai as a moodboard | hosted Krea2ImageNode |
moodboard UUIDs exist only on their service |
| Style from 1 to 10 reference images, no local weights | hosted, chain Krea2StyleReferenceNode |
up to 10 refs, each with its own strength |
| Same style locally, one reference | local Krea 2 Turbo + krea2_style_reference LoRA |
free, no cap at 1K, see MODELS.md |
| Output above 1K | local only | the hosted node's resolution combo has exactly one option, 1K |
| Photoreal stills, opinionated photographic look, low VRAM | FLUX.1 Krea Dev (graph below) | 11.90 GB (11.09 GiB) fp8, ordinary FLUX graph |
| Instruction editing that keeps a face | local Krea 2 + comfyui-krea2edit |
see the ecosystem map below |
| Depth / canny / pose control on Krea 2 | local Krea 2 + comfyui-krea2-controlnet |
core has no Krea 2 ControlNet |
| Realtime or streaming video, interactive prompt changes | their own server, DiffSynth-Studio or LightX2V | the one ComfyUI pack has 0 stars and is unproven, see below |
What ships with it
1 file 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.
- 6d ago First seen · 160 lines · 104 tokens per session scan A bf08a5a72186
krea is a skill published in the GitHub repository SlavaSexton/ComfyUI-Agent-Kit (93 stars, last pushed 3d ago), licensed Apache-2.0. It adds 104 tokens to every session and 3,091 once invoked, about $0.0005 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
regional-culture-poster
Convert Chinese provinces, municipalities, autonomous regions, cities, counties, ancient cities, and cultural regions into restrained contemporary posters led by one culturally transformed Chinese character. Supports direct one-shot final image generation and layered production. Use when the user asks for a…
guizang-social-card-skill
Generate Guizang-style social card image sets, Live Photo motion cards, material-first Live Photo puzzle layouts, triple Live Photo collages, long-video-to-Live-Photo treatments, and WeChat official account cover pairs from articles, scripts, screenshots, product notes, subtitles, photos, or user-supplied videos. Use…
minimax-h3
Write, debug and structure prompts for MiniMax H3 video generation (T2VA, I2VA, FL2VA, L2VA, Ref2VA) and configure its ComfyUI workflow. Use when the user mentions MiniMax H3, minimaxh3, fl2va, ref2va, MiniMaxH3ReferenceToVideo, reference-to-video, asks to animate a photo, write a video prompt, pick a model quant, or…
promo-video
Design and render a premium, beat-locked product video with Remotion — an ORIGINAL template every time, themed to the product. Intake-first — ask for the website URL, research it, then ask a question batch (duration, theme, scenes, focus, audience…) so the requester never has to explain their product. Use when the…
ai-short-drama
精细化 AI 短剧 IP 创作技能（v0.6.0）。三阶段架构：Phase 1 创作（剧本+ref图，反复迭代）→ Phase 1.5 分镜图（每 grid 1-4 张候选静态图 = 视频首帧，工业级核心层）→ Phase 2 出片（按集解锁，4 模自动选）。v0.6.0 关键升级：① ref 库工艺偏置铁律（现代摩天楼易拟物，古建筑必出 chibi 人体，前期 IP 设计阶段就要避坑）② 即梦 5.0 失败模式 + 敏感词清单（3 类 fail 区分 / prompt 1500 字硬上限 / 反派词替换表 / 暧昧词清单）③ 分镜图 8 段 prompt…
cinematic-frame-director
Turn a user's idea, scene description, or mood into ONE production-grade cinematic still-image prompt in English for text-to-image models (Seedream, Midjourney, Flux, DALL-E, or any image generator). Use this skill whenever the user asks for an image prompt, a cinematic frame, a "movie still" look, a photorealistic…