content-generation-skills

mcp
Guvenlik Denetimi
Uyari
Health Uyari
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 5 GitHub stars
Code Uyari
  • process.env — Environment variable access in scripts/export-visual-guides.mjs
  • process.env — Environment variable access in scripts/sync-mcp-snapshot.mjs
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

Mira AI content generation skills: genre playbooks and prompting knowledge for AI agents that generate video and images through the Mira MCP server

README.md

Mira skills

Mira skills: content generation skills for AI agents

License: MIT
Version
Skills
Agent Skills
check

Skills for AI agents that generate images, video and 3D on Mira AI. Ten of them are playbooks: UGC, product, ads, cinematic, explainer, anime, illustration, 3D toon, pixel art, 3D model. Pass a playbook id as the skill parameter of generate_video, generate_image or generate_3d and the platform's prompt pipeline applies the genre's doctrine itself. Five are knowledge: how to drive the Mira MCP server, how the fourteen video models differ, how the image models differ, the camera, light, look, rhythm and continuity banks the pipeline is built on, and how to build and shoot scenes in the user's own Blender through the bridge tools.

Works with Claude Code, Codex, Cursor and any agent that reads SKILL.md files. Generation runs through the Mira MCP server at https://mcp.mybots.pro/mcp, signs in with your Mira account and spends your own credits. No API keys.

Install

npx skills add Global-AI-Group-Inc/content-generation-skills

Claude Code marketplace:

/plugin marketplace add Global-AI-Group-Inc/content-generation-skills
/plugin install mira@mira

Then connect the MCP server in your agent. The page at mira.mybots.pro/mcp has the exact config for each client.

What is inside

Skill Kind What it does
mira-blender-scene core Build and shoot scenes in the user's own Blender through the blender_* tools: blockout in metres, lights, camera, keyframes, generated assets, playblast → video model.
mira-generate core Drive the Mira MCP tools: connect, pick a model, attach brand assets, generate, wait, extend, upscale, build 3D and rig it. Explains what the pipeline keeps and what it fills in.
mira-video-prompting knowledge Which of the fourteen video models to pick and what each one actually does with a prompt. One reference file per recommended model.
mira-image-prompting knowledge The image models and the craft banks for stills: light, optics, composition, grade, materials, and the words that hurt.
mira-craft knowledge The camera, light, look, rhythm and continuity banks the pipeline reads. Generated from the live registry.
mira-ugc playbook Phone-shot creator content: review, unboxing, how-to, testimonial.
mira-product playbook The product as the hero: studio, lifestyle, hands.
mira-ads playbook Short paid creative with a hook in the first beat and a clean closing frame.
mira-cinematic playbook One strong image, film light and optics.
mira-explainer playbook Show how it works in plain light and plain framing.
mira-anime playbook Hand-drawn animation, not footage.
mira-illustration playbook Drawn and painted artwork in motion.
mira-toon3d playbook Stylised 3D animation and CGI key visuals.
mira-pixel playbook Pixel art on a visible grid.
mira-3d playbook One object as a 3D model: text or photos → GLB/FBX/OBJ/USDZ, optional PBR and a rigged humanoid with animation clips.

Playbook ids are the last part of the skill name: mira-ugc is skill: "ugc".

How the playbooks work

The platform keeps every decision you state in the prompt and fills only the craft you left out. A playbook adds one more layer: the doctrine of a genre goes into the planning step of the pipeline, so a UGC clip gets handheld framing, window light and imperfect texture without you writing any of it. You describe the subject and the moment. The playbook and the pipeline do the rest.

Each playbook also carries a short interview: two or three questions with fixed options. In an agent, ask them in chat when the answer is not obvious. In the Mira composer they show up as quick-answer chips. Skip a question and the default applies.

Repository layout

skills/<name>/SKILL.md       the router; short on purpose
skills/<name>/references/    the detail, read one file at a time
catalog.json                 index used by Mira's own agent and by the MCP server
scripts/check.mjs            frontmatter, line limits, banned words, bank keys
scripts/build-catalog.mjs    regenerates catalog.json
scripts/export-visual-guides.mjs   copies skills to Mira's shelf
scripts/sync-mcp-snapshot.mjs      copies skills into the MCP server build
VERSION                      bump on every content change

Files marked generated from the registry are written by build_skills.py in the platform repo. Edit the registry, not the file.

Contributing

Open a pull request. Keep SKILL.md under 150 lines, write in English, and run node scripts/check.mjs before pushing. Doctrines must not contain bank keys such as push_in or rembrandt: those are for choosing, the prompt gets the described light or move in plain words.

License

MIT, copyright Global AI Group, Inc. (the company behind Mira AI). Attribution is welcome, not required.

Yorumlar (0)

Sonuc bulunamadi