The studio's source of truth — every model · workflow · skill · MCP tool · repo · framework · specialist we run, with freshness/health provenance. We didn't buy the stack; we cataloged + control it. Source: temerarii/catalog/*.yaml.
| Item | Domain | What | Health |
|---|---|---|---|
| Cloudinary — media delivery + transformation CDN cloudinary | ops delivery | Programmable media CDN that closes the missing transform/delivery leg between asset generation (fal/kie/replicate/Remotion/HyperFrames/ffmpeg) and publishing. One master upload -> every social aspect ratio, format (WebP/AVIF), and size v | planned unverified .mcp.json (asset-management-mcp) + pip cloudinary-cli |
| Box — governed content archive + client handoff + AI source ingestion box | ops archive | Enterprise content layer at the cold/governed end of the pipeline (NOT the hot publish path). Permissioned, audit-logged drop point for client deliverables, and via mcp-server-box-remote + box-for-ai an AI-readable source of client-provided | planned unverified .mcp.json (mcp-server-box-remote) + pip box-python-sdk |
| Higgsfield — multi-model image/video generation (OAuth MCP) higgsfield | creative generation | One OAuth-authenticated MCP server exposing 30+ generation models (Seedance 2.0, Sora 2, Veo, Kling, Nano-Banana-Pro, Flux) plus Higgsfield Soul (face-faithful character identity) and Cinema Studio (multi-shot, one consistent look), with bu | planned unverified claude mcp add --transport http --scope user higgsfield https://mcp.higgsfield.ai/mcp (OAuth) + npx skills add higgsfield-ai/skills |
| Item | Domain | What | Health |
|---|---|---|---|
| cloudinary/* — SDKs, CLI, transformation + MCP servers cloudinary-repos | ops delivery | Cloudinary's open-source surface backing the delivery/transform tool. Adopt-worthy: asset-management-mcp / structured-metadata-mcp / environment-config-mcp (first-party MCP servers, the agentic path), cloudinary-cli (Python, batch ops), ass | planned unverified npm / pip per package |
| box/* — SDKs, CLI, MCP server + AI toolkit box-repos | ops archive | Box's open-source surface backing the archive/handoff tool. Adopt-worthy: mcp-server-box-remote (secure agent access to enterprise content), box-for-ai (developer AI toolkit), box-python-sdk / box-node-sdk + boxcli (scripted handoff from en | planned unverified npm / pip per package |
| mrdoob/three.js threejs | creative 3d | Canonical three.js repo — source-of-truth + version pin for R3F (Lab) + @remotion/three + 3d-force-graph. | wired current ● live |
| nytimes/three-loader-3dtiles three-loader-3dtiles | creative 3d | Streams Cesium/Google 3D Tiles into three.js -> Lab geospatial/photoreal-city experiments. | planned unverified |
| WICG/html-in-canvas html-in-canvas | engineering rendering | Render live, interactive, accessible DOM as WebGL/WebGPU textures. Production-viable via HyperFrames. | planned flagged |
| microsoft/markitdown markitdown | engineering ingestion | Anything -> Markdown (PDF/Office/images-OCR/audio/HTML/CSV/YouTube). The ingestion leg of the content spine. | planned current |
| zarazhangrui/frontend-slides frontend-slides | creative design | MD/PPTX -> single-file HTML decks, 34 templates. A markdown->HTML leg (qa_gated). | planned current |
| microsoft/playwright-cli playwright-cli | engineering browser | Engine behind browser-MCP inspiration sourcing + web QA. | wired current ● live |
| obsidianmd/obsidian-headless (+ obsidian-api) obsidian-headless | engineering memory | Headless Obsidian for syncing/rendering the L2 vault -> JSON export -> 3d-force-graph. | planned unverified |
| vasturiano/3d-force-graph 3d-force-graph | creative 3d | three.js-based 3D force graph — renders the Part-14 interactive registry universe from the catalog. | planned current |
| getzep/graphiti graphiti | engineering memory | Temporal knowledge graph for the DEFERRED L3 runtime memory ("what changed when" + impact queries). | planned unverified |
| neo4j-contrib/neovis.js neovis-js | engineering memory | Neo4j graph viz for the optional L3 Query/Timeline tabs (deferred). | planned unverified |
| mem0ai/mem0 mem0 | engineering memory | Agent memory layer — candidate for L3 runtime recall (deferred). | planned unverified |
| vercel-labs/skills vercel-labs-skills | agentic harness | The `npx skills add` installer (GitHub-as-registry) — how vendor skill-packs land in .claude/skills/. | wired current ● live |
| facebookresearch/sam-3d-objects sam-3d-objects | creative 3d | Single photo -> 3D (Gaussian splats + mesh). Real-object reconstruction; splats -> three.js/R3F. | planned flagged |
| facebookresearch/sam2 sam2 | creative edit | Image+VIDEO segmentation -> subject isolation/rotoscope for buttercut + ffmpeg + Lab. Adopt-now. | planned current |
| facebookresearch/faiss faiss | engineering memory | Vector search -> memory/RAG L3 + visual asset library. Adopt-now (must-wire with DINOv2). | planned current |
| facebookresearch/demucs demucs | creative music | Music stem separation -> remix/score editing. Adopt-now. | planned current |
| facebookresearch/dinov2 dinov2 | engineering memory | Visual embeddings -> asset search/dedup (pairs with faiss). Adopt-now. | planned current |
| awslabs/gluonts gluonts | ops analytics | Probabilistic forecasting -> predict marketing metrics (Performance pillar). Adopt-now. | planned current |
| awslabs/git-secrets git-secrets | engineering security | Secret-scan pre-commit -> pairs with the AI-security gate. Adopt-now. | planned current |
| googlecreativelab/creatability-components creatability-components | engineering accessibility | Accessible creative web components -> the accessible-Lab + brand-book Accessibility part. | planned current |
| XLabs-AI/x-flux (+ x-flux-comfyui) x-flux | creative image | ControlNet + IP-Adapter + LoRA train/infer for FLUX -> brand-consistent controllable image gen. | planned flagged |
| microsoft/AI-Red-Teaming-Playground-Labs ai-red-teaming-labs | engineering security | 12 challenges (incl. indirect prompt injection) + PyRIT -> backs the AI-security gate. | planned current |
| daviddinkevich/Story2Board story2board | creative 3d | Training-free, identity+layout-preserving storyboard technique -> informs Higgsfield Soul-ID prompts. | planned current |
| joenorton/comfyui-mcp-server (+ SaladTechnologies/comfyui-api) comfyui-mcp-server | creative image | ComfyUI control — MCP bridge for agentic calls; Salad comfyui-api for stateless cloud-GPU batch. | planned unverified |
| ahujasid/blender-mcp blender-mcp | creative 3d | Canonical Blender MCP bridge — connects Blender to Claude via a socket addon for AI-assisted 3D: scene/object create-delete-modify, material mgmt, Python execution in Blender, viewport screenshots, and asset download (Poly Haven, Hyper3D Ro | planned current v1.5.5 |
| 8beeeaaat/touchdesigner-mcp (+ bottobot/td-mcp) touchdesigner-mcp | creative 3d | TD live control via WebServer-DAT .tox (:9981) + bottobot docs/introspection. Realtime/generative visuals for the Lab. | planned unverified |
| splinetool (r3f-spline) spline | creative 3d | Spline 3D scenes embedded via r3f-spline -> Lab hero / interactive 3D; community scenes export to three.js. | planned unverified |
| Item | Domain | What | Health |
|---|---|---|---|
| Harness — goal -> stages -> gates harness-goal-stages-gates | agentic harness | The ONE workflow (engine/goal.py): a goal cascades through STAGES (8, do the work) proven by GATES (8 + qa_gate, define "aligned"). Graceful-empty-pass: an absent artifact is a NO-OP pass, so `goal.py status` is green today and ga | wired current ● live |
| Spec-Kit (.specify) spec-kit | agentic harness | speckit.* constitution/specify/plan/tasks/implement/analyze — spec-driven build discipline. | wired current ● live |
| Single Unified Scaffold single-scaffold | agentic architecture | temerarii-studio IS the one canonical scaffold — .specify + engine/ + brands/ + temerarii/catalog/ + apps/office/ + governance/ + docs/ wired so one change (a token, gate, calendar campaign, registry entry, tool's health) propagates to ever | wired current ● live |
| 4-4-5 Calendar + drift gate calendar-445 | ops planning | 4-4-5 fiscal calendar with parallel lanes (track_of) + a drift gate keeping artifacts week-consistent. | wired current ● live |
| Brand Voice Rails brand-voice-rails | ops brand | governance/brand.json voice rails (pragmatic/"Minus the Hype" + 29 banned words + frames), enforced by qa_gate on all output. The brand's discipline layer for copy. | wired current ● live |
| Brand Motion Rails (MOTION-SACRED) brand-motion-rails | creative brand | 7 sanctioned moves (kinetic-build, mark-draw, crossfade-8f, receipts-count, ember-drift, spatial-parallax R3F, cinematic-push) + forbidden list + reduced-motion poster. Motion is sacred — kinetic typography / three.js / Spline / scroll-acti | wired current ● live |
| 9-Output Social/Video Matrix nine-output-matrix | creative video | One master render -> every platform format/aspect/size (Remotion today; HyperFrames as the HTML-native alt). | wired current ● live |
| HTML-native, agent-authored, no-build substrate html-native-substrate | engineering architecture | Per Simon Willison — HTML is the universal artifact substrate. office + brand book (single-file HTML), HyperFrames (HTML->MP4), html-in-canvas (HTML->WebGL texture), cinematic-site-components, blog-automation (HTML->5 CMS). Prefer | wired current ● live |
| Research -> Markdown -> HTML -> Publish (content spine) research-md-html-publish | engineering pipeline | RESEARCH (Perplexity/Tavily/Exa/Apify) -> ANYTHING->MARKDOWN (MarkItDown) -> KNOWLEDGE (Obsidian vault) -> MARKDOWN->HTML (frontend-slides / Chef Sumi / brand renderer, qa_gated) -> PUBLISH (HyperFrames / blog-automation / | planned current |
| Hybrid Memory (L1/L2/L3) hybrid-memory | engineering memory | L1 machine truth = typed YAML + Pydantic + gates (temerarii/catalog/, this registry). L2 human graph = generated Obsidian vault ([[links]], git-native). L3 runtime memory = graph+vector (Graphiti/Mem0/Neo4j+neovis/faiss) DEFERRED behind a & | wired current ● live |
| Tool-Health & Freshness Audit tool-health-audit | ops governance | Every tool/skill-pack/repo we add is versioned, sourced, licensed, and audited; deprecated/ flagged-without-ack can't ship. Checks latest version, deprecation/archived, flag-risk (proxied/unofficial APIs), key validity, MCP config drift. St | planned current |
| Design-Inspiration Sourcing design-inspiration-sourcing | creative design | sample (WebFetch / Browser MCP / 21st.dev) -> brand-rail filter (dark/ember/Space-Grotesk, reject forbidden motion) -> annotated shortlist -> feeds Lab + brand-book Parts 02-09. 12 sources (godly/awwwards/21st.dev/lapa/landingfolio | wired current ● live |
| Observability — OpenTelemetry observability-otel | ops observability | Trace the swarm + harness (run=trace, MCP call=span w/ latency+cost+token attrs via GenAI semantic conventions). Feeds the Tool-Health audit + cost-tracker. Earn-its-keep: start zero-infra (file/console exporter), add a collector only when | planned current |
| AI-Security Gate (indirect-injection hardening) ai-security-gate | engineering security | We feed external research/OSINT/scraped content into tool-enabled agents = classic indirect prompt-injection risk. Harden via Microsoft AI-Red-Teaming-Playground-Labs + PyRIT; pairs with cso/mcp-security/BlueRock MCP monitoring. | planned current |
| Item | Domain | What | Health |
|---|---|---|---|
| Chef Sumi (phazurlabs/sumi) chef-sumi | creative design | The design DISCIPLINE engine — 34 commands across Make (/style /tokens /palette /type /component /screen /dark /responsive), Review (/grade /audit /roast /a11y /qa), Plan (/brief /research /benchmark /measure). Drives governance/brand.json | wired current ● live v9.0.0 |
| fal-ai-community/skills (genmedia) fal-ai-skills | creative generation | genmedia CLI + Commercial/Marketing/Cinematography skills, 10 modalities, models catalog + recipes. | planned current ● live npx skills add fal-ai-community/skills; genmedia init |
| higgsfield-ai/skills higgsfield-skills | creative generation | /higgsfield:generate|soul-id|product-photoshoot|marketplace-cards — the skill layer over the OAuth MCP. | planned unverified npx skills add higgsfield-ai/skills |
| remotion-dev/skills remotion-skills | creative video | Official Remotion skills pack. | planned unverified npx skills add remotion-dev/skills (DEFER) |
| CloudAI-X/threejs-skills threejs-skills | engineering 3d | Vendor three.js skill pack for R3F/WebGPU/TSL workflows. | planned unverified npx skills add CloudAI-X/threejs-skills |
| heygen-com/hyperframes hyperframes | creative video | HTML/CSS/animation -> deterministic MP4 (headless Chrome + FFmpeg). HTML-native (no React), agent-native; supports GSAP/Lottie/Three.js/Anime.js + component catalog + Lambda render. Auto-enables CanvasDrawElement at render time (html-in- | planned current ● live v0.6.64 |
| robonuggets/hyperframes-helper robonuggets-hyperframes-helper | creative video | 3-level HyperFrames workflow, 16 error patterns, HTML/CSS/GSAP recipes, storyboard templates. | planned current ● live copy to .claude/skills/ |
| robonuggets/seedance-skill robonuggets-seedance-skill | creative video | Seedance 2.0 via fal — screenshot/prompt -> liquid-glass app promo reels (4-10s). | planned current ● live copy to .claude/skills/ |
| robonuggets/cinematic-site-components robonuggets-cinematic-site-components | creative design | 30 single-file HTML/CSS/GSAP modules (scroll reveals, parallax, magnetic, mesh gradients; no build). | planned current ● live copy to .claude/skills/ |
| robonuggets/higgsfield-skill robonuggets-higgsfield-skill | creative generation | OAuth-MCP path to Higgsfield (30+ models) — the gap-closer that activates Higgsfield without a raw key. | planned unverified copy to .claude/skills/ + claude mcp add http higgsfield (OAuth) |
| robonuggets/marp-slides robonuggets-marp-slides | creative design | Markdown -> decks (pitch / thought-leadership). Decks only — NOT the brand book (that stays HTML). | planned current ● live copy to .claude/skills/ |
| robonuggets/design-system robonuggets-design-system | creative design | Design-system skill — pairs with Chef Sumi for brand-book Parts 02-09. | planned current ● live copy to .claude/skills/ |
| robonuggets/plan-for-goal robonuggets-plan-for-goal | agentic harness | brief -> /goal pattern; a useful planning pattern next to our engine/goal.py harness. | planned current ● live copy to .claude/skills/ |
| robonuggets/agentic-ai-starter-kit robonuggets-agentic-ai-starter-kit | agentic harness | CLAUDE.md / settings scaffolding patterns for agentic projects. | planned current ● live copy to .claude/skills/ |
| robonuggets/markitdown-guide robonuggets-markitdown-guide | engineering ingestion | Usage skill for MarkItDown (any file -> Markdown) feeding the research->vault ingestion leg. | planned current ● live copy to .claude/skills/ |
| kepano/obsidian-skills kepano-obsidian-skills | engineering memory | Obsidian skills for the L2 knowledge-graph vault (the human view of the registry). | planned unverified npx skills add kepano/obsidian-skills |
| anthropics/skills anthropics-skills | agentic harness | First-party Anthropic skills pack. | planned unverified npx skills add anthropics/skills |
| Item | Domain | What | Health |
|---|---|---|---|
| Avatar Clone Pipeline avatar-clone | creative avatar | HeyGen + ElevenLabs avatar-video pipeline with hardcoded studio identities. | wired current ● live |
| Buttercut Video Editing buttercut-editing | creative edit | AI video editing (rough-cut / edit-point detection) built on Remotion. | wired current ● live |
| Cinematic Prompts cinematic-prompts | creative image | Prompt scaffolds for cinematic image/video generation (shot/lighting/lens language). | wired current ● live |
| Content Automation content-automation | agentic show | Overview skill for the end-to-end content pipeline (research -> script -> media -> publish). | wired current ● live |
| Multi-Platform Distribution distribution | ops distribution | Blotato 16-account distribution across 2 profiles (video -> YouTube only; social gets derived cuts). | wired flagged |
| FFmpeg Mastery ffmpeg-mastery | engineering edit | Comprehensive FFmpeg reference — encode/convert/composite, NVENC GPU path on RTX 4070. | wired current ● live ffmpeg 8.0.1 |
| Flow Transformer flow-transformer | agentic show | Knowledge base for analyzing n8n flows and mapping them to Claude Code + MCP tools. | wired current ● live |
| Image Generation image-generation | creative image | PiAPI/Replicate/Flux image workflows (palette-aware, brand-railed). | wired current ● live |
| AI Model Price Matrix price-matrix | ops analytics | Per-model cost reference for budget-aware model routing across providers. | wired current ● live |
| Remotion Mastery remotion-mastery | creative video | Deep Remotion technical knowledge for programmatic React video + the 9-output matrix. | wired current ● live v4.0.399 |
| Research Pipeline research-pipeline | agentic research | Apify/Perplexity/HN/Reddit parallel research + curation for viral-news angles. | wired current ● live |
| Video Pipeline video-pipeline | engineering video | FFmpeg-based video processing / format conversion / compositing stage helpers. | wired current ● live |
| Voice Input System voice-input | agentic voice | Voice -> Claude -> content pipeline (Win+H dictation; Whisper transcription). | wired current ● live |
| Build Everything build-everything | agentic show | Standing build system — /build-everything <target> builds any new thing (campaign, app, case study, engine feature) the disciplined way: deep-research preflight -> spec -> build -> gates. | wired current ● live |
| Remotion Best Practices remotion-best-practices | creative video | Best-practices skill for Remotion video creation in React (pairs with remotion-mastery). | wired current ● live |
| Item | Domain | What | Health |
|---|---|---|---|
| kling-2.5 kling-2.5 | video | ||
| runway-gen4 runway-gen4 | video | ||
| veo-3 veo-3 | video | ||
| sora-2 sora-2 | video | ||
| pika-2.2 pika-2.2 | video | ||
| higgsfield higgsfield | video | ||
| heygen-avatar heygen-avatar | avatar | ||
| flux-1.1-pro-ultra flux-1.1-pro-ultra | image | ||
| flux-kontext flux-kontext | image | ||
| z-image-turbo z-image-turbo | image | ||
| midjourney-v7 midjourney-v7 | image | ||
| elevenlabs-multilingual-v2 elevenlabs-multilingual-v2 | voice | ||
| qwen3-tts qwen3-tts | voice | ||
| openai-whisper openai-whisper | voice | ||
| suno-v4.5 suno-v4.5 | music | ||
| udio-1.5 udio-1.5 | music | ||
| musicgen musicgen | music | ||
| perplexity-sonar perplexity-sonar | research | ||
| tavily-search tavily-search | research | ||
| apify-actors apify-actors | research | ||
| exa-search exa-search | research | ||
| ffmpeg ffmpeg | edit | ||
| remotion remotion | edit | ||
| comfyui comfyui | edit | ||
| blender blender | 3d | ||
| blotato blotato | distribution | ||
| cloudinary cloudinary | distribution | ||
| ga4-data-api ga4-data-api | analytics | ||
| posthog posthog | analytics |
| Item | Domain | What | Health |
|---|---|---|---|
| AI Clone Automation ai-clone-automation | avatar | End-to-end AI clone avatar automation (clone -> script -> render -> post). | |
| AI Social Video Generator ai-social-video-generator | video | Generates short social videos from a topic/prompt for the distribution matrix. | |
| Blotato Post Everywhere blotato-post-everywhere | distribution | Fan-out a finished asset to all connected Blotato accounts (video -> YouTube only). | |
| HeyGen Clone heygen-clone | avatar | HeyGen avatar-clone video generation with the studio avatar identity. | |
| Temerarii AI Avatar — Trending News temerarii-ai-avatar-trending-news | avatar | Research trending news -> script -> avatar VO -> render -> post (the flagship pipeline). | |
| UGC Nano-Banana ugc-nanobanana | image | UGC-style image/video generation via Nano-Banana (Gemini image) for product/social. | |
| Veo3 + Blotato Autopost veo3-blotato-autopost | video | Generate video with Veo 3 then auto-post via Blotato. | |
| Veo3 Generation veo3-in-n8n | video | Veo 3 text/image -> video generation (the gen core of the autopost flow). | |
| Viral News AI Avatar viral-news-ai-avatar | avatar | Viral-news research -> selection -> script -> avatar video (active migration in flows/active). | |
| Generate Image (dual-path) generate-image | image | Image generation via workflows/z-image — fal primary, HuggingFace free fallback. | |
| Generate Music generate-music | music | Music/score generation via workflows/music (MusicGen on HuggingFace). | |
| Composite Video (FFmpeg) composite-video | edit | GPU video compositing/encode via workflows/ffmpeg_compositor + generators/video/ffmpeg. | |
| Render Storyboards (Remotion 9-output) render-storyboards | video | Remotion render of the storyboard registry into the 9-output social/video matrix. | |
| Distribute distribute | distribution | Push finished assets to channels via engine/stages/distribute.py (Blotato). | |
| Syndicate Blog syndicate-blog | distribution | Cross-post written content (Blogger/Reddit/Medium today; Duda+4 CMS via cms-publisher planned). | |
| Sync Insights sync-insights | research | Live sitemap -> catalog sync + merge + used-rotation for the Insights corpus. |
| Item | Domain | What | Health |
|---|---|---|---|
| Screenwriter screenwriter | creative development | Writes scripts, monologues, and dialogue in the studio voice. | wired current ● live |
| Story Architect story-architect | creative development | Structures narrative arcs, beats, and act structure for content. | wired current ● live |
| World Builder world-builder | creative development | Builds consistent settings, lore, and visual worlds for campaigns. | wired current ● live |
| Casting Director casting-director | creative preproduction | Selects avatars/voices and casts on-screen talent for productions. | wired current ● live |
| Concept Artist concept-artist | creative preproduction | Generates concept/key art and mood direction (fal/Higgsfield/Flux). | wired current ● live |
| Production Designer production-designer | creative preproduction | Defines the visual design language and set/scene aesthetics for a piece. | wired current ● live |
| Storyboard Artist storyboard-artist | creative preproduction | Turns scripts into shotlists/storyboard frames (Higgsfield Soul ID + storyboard-registry). | wired current ● live |
| 3D Supervisor 3d-supervisor | creative production | Oversees 3D asset gen + scene assembly (Blender/three.js/Meshy/TRELLIS). | wired current ● live |
| Animation Director animation-director | creative production | Directs motion/animation timing and performance across shots. | wired current ● live |
| Cinematographer (Director of Photography) cinematographer | creative production | Controls camera, lensing, and lighting language for generated footage. | wired current ● live |
| Motion Graphics Artist motion-graphics-artist | creative production | Kinetic typography, mark-draw, ember-drift — the sacred motion layer (Remotion/HyperFrames/CSS). | wired current ● live |
| Voice Director voice-director | creative production | Directs VO performance, casting, and read style (ElevenLabs voice ID). | wired current ● live |
| Colorist colorist | creative postproduction | Color grade + look development to the brand palette (navy/ember/near-black). | wired current ● live |
| Editor editor | creative postproduction | Assembles cuts, pacing, and edit-point decisions (buttercut + ffmpeg). | wired current ● live |
| Music Supervisor / Composer music-supervisor | creative postproduction | Sources/generates score + music beds (Suno/Udio/MusicGen) and syncs to picture. | wired current ● live |
| Sound Designer / Audio Director sound-designer | creative postproduction | SFX, ambience, and audio mix; premix + loudness alignment. | wired current ● live |
| VFX Compositor vfx-compositor | creative postproduction | Compositing, keying, and visual effects integration (ffmpeg/SAM2 rotoscope). | wired current ● live |
| Analytics Specialist analytics-specialist | ops distribution | Measures performance, attribution, and forecasting (Performance pillar; gluonts). | wired current ● live |
| Marketing Specialist marketing-specialist | ops distribution | Positioning, messaging, and campaign strategy in the "Minus the Hype" voice. | wired current ● live |
| Platform Specialist platform-specialist | ops distribution | Per-platform format/spec optimization for the 9-output distribution matrix. | wired current ● live |