Agent Skills

milvus-meme-sticker

imagezc2775841212.1K installs

Create tiny no-text Milvus-style sticker memes for technical marketing, WeChat-style article moments, community posts, and developer-group reactions. Use when the user wants an original meme/sticker with an abstract Milvus bird/eagle mascot, strong exaggerated emotion, internet-native sticker energy, light-blue accents without making the whole character blue, and deterministic small-size exports such as 512, 240, 120, or 50 px.

Install

npx skills add https://github.com/zc277584121/marketing-skills --skill milvus-meme-sticker
SKILL.md

Milvus Meme Sticker

Create original sticker-style memes for technical marketing moments. The output should feel like something people would send in a developer WeChat group: compact, expressive, funny, emotionally resonant, and safe to publish because it does not reuse existing meme templates, celebrities, screenshots, anime, or other IP.

Read references/style-guide.md before generating the first sticker in a session.

Workflow

  1. Identify the emotion from the source paragraph or user request: overwhelmed, confused, relieved, proud, shocked, exhausted, or "why is this so hard?".
  2. Generate three square variants first unless the user asks for one exact direction.
  3. Use an abstract Milvus bird/eagle mascot as the recurring character, not a logo copy.
  4. Keep the image text-free. No captions, labels, letters, UI words, fake code snippets, or speech bubbles with text.
  5. Prefer a compact sticker composition with a clear silhouette and large readable facial expression.
  6. Save the best or selected source image at high resolution, then run scripts/resize_sticker.py to export true sticker sizes.
  7. Insert or deliver the small exported image, not only the large generation.

Prompt Rules

Use prompts like this, adapting only the emotion and scene:

Square 1:1 sticker-style meme illustration, no text, no letters, no captions.
Original abstract mascot inspired by a Milvus-like bird/eagle, simplified and
not copying any exact logo. The character has a light body with only partial
soft blue accents, a non-blue beak and mouth, oversized expressive eyes, and a
funny exaggerated [EMOTION] expression.

Scene: [SCENE]. Developer-group chat sticker vibe, hand-drawn, thick clean
outline, white sticker border, compact composition readable at small size.
Use a light-blue accent palette with colorful secondary accents; do not make
the whole character blue. No real brands, no existing meme templates, no
celebrities, no screenshots, no IP characters, no text anywhere.

Style Guardrails

  • Keep the mascot abstract: bird/eagle-inspired, rounded and sticker-like, not a realistic animal and not a copy of any official mark.
  • Use light blue as an accent, not as a full-body flood. Avoid blue teeth, blue mouth interiors, or an all-blue face that loses expression.
  • Make the beak, mouth, eyes, sweat, and emotion marks clearly separated in color and value.
  • Favor expressive face and body posture over detailed props.
  • Use props only when they explain the pain point: documents, connector plugs, database cylinders, cloud buckets, chat bubbles, tickets, code brackets.
  • Avoid text because this sticker must travel across languages and stay legible at tiny sizes.
  • Avoid photorealism, 3D toy rendering, glossy mascot branding, and polished corporate illustration. It should feel like a hand-drawn sticker meme.

Small-Size Export

Generated images are usually too large to feel like real stickers. Always create small derivatives after generation:

python3 scripts/resize_sticker.py path/to/source.png --out-dir path/to/output-dir

Defaults create:

  • 512x512 for article insertion when some detail should remain visible.
  • 240x240 for WeChat-style sticker main image scale.
  • 120x120 for thumbnail checks.
  • 50x50 for tiny panel legibility checks.

Use the smallest version that still reads in context. For Markdown articles, prefer 240x240 or 512x512 plus explicit display sizing if needed.

Quality Check

Before showing the final result, inspect the image and verify:

  • No visible text or pseudo-text.
  • The mascot is not entirely blue.
  • The mouth, beak, eyes, and expression are readable.
  • The emotion is obvious even at 240x240.
  • The sticker silhouette is compact and not too busy.
  • The image does not resemble a known meme template, celebrity, anime, or existing IP character.

Related skills

ai-image-generationgenmedia-labs713KGenerate and edit images on RunComfy via the `runcomfy` CLI — a smart router across the full image-model catalog: FLUX 2 (Klein 9B/4B, Pro, Dev, Flash, Turbo, Max), Google Nano Banana 2 / Pro, OpenAI GPT Image 2, ByteDance Seedream 5 / 4-5 / 4-0 and Dreamina 4-0, Alibaba Qwen Image and Z-Image Turbo, Wan 2-7. Covers both text-to-image (t2i) and image-to-image / edit (i2i) endpoints — the skill picks the right model for the user's actual intent (typography precision, photoreal portraits, sub-secoai-image-generation101-skills547KGenerate AI images with GPT-Image-2, FLUX, Gemini, Grok, Seedream, Reve and 50+ models via inference.sh CLI. Models: GPT-Image-2, FLUX Dev LoRA, FLUX.2 Klein LoRA, Gemini 3 Pro Image, Grok Imagine, Seedream 4.5, Reve, ImagineArt. Capabilities: text-to-image, image-to-image, inpainting, LoRA, image editing, upscaling, text rendering. Use for: AI art, product mockups, concept art, social media graphics, marketing visuals, illustrations. Triggers: flux, image generation, ai image, text to image, stnano-banana-2prime-skills424KGenerate images with Google Nano Banana 2 (Gemini-family flash-tier text-to-image) on RunComfy — bundled with the model's documented prompting patterns so the skill gets sharper output than naive prompting against the same model. Documents Nano Banana 2's strengths (rapid iteration, in-image typography rendering, predictable framing, optional web-grounded context), the resolution-tier pricing, the safety-tolerance dial, and when to route to Nano Banana Pro / GPT Image 2 / Flux 2 / Seedream insteimage-editprime-skills424KEdit images on RunComfy — this skill is a smart router that matches the user's intent to the right edit model in the RunComfy catalog. Picks Nano Banana Edit (batch up to 20, identity-preserving default), OpenAI GPT Image 2 Edit (multilingual in-image text rewrite, multi-ref composition, layout precision), Flux Kontext Pro (single-ref high-fidelity local edit), or Z-Image Turbo Inpaint (mask-driven precise region edit). Bundles each model's documented prompting patterns so the skill gets sharper

Search skills and MCP servers

Fuzzy search across 23,137 skills and servers