theorems-harness
claudecodexTheorem's Harness: GraphQL-first MCP for adaptive planning, context, coordination, execution, validation, memory, and learning.
37 skills · 11 commands · 10 agents · 1 MCP server · hooks
Makes Claude Code genuinely good at designing knowledge-graph answer experiences with cosmos.gl on top of DuckDB-WASM, Mosaic, and vgplot. Owns the SceneDirective adapter, the three-picker ControlDock (Position, Weight, Edges), and the recipe library that turns novel ideas into usable images.
Plugin cosmos-pro from marketplace codex-marketplace in Travis-Gilbert/claude-marketplace. Commands read from the repo's manifest — 1 month ago.
cosmos-designVisual planning partner for cosmos.gl + Mosaic + DuckDB-WASM + vgplot. Use when planning knowledge-graph answer experiences, choosing layer compositions, designing the Position/Weight/Edges picker UI, planning construction sequences (Galaxy, Filter, Build, Crystallize, Explore), or deciding between
cosmos-foundationsCanonical cosmos.gl v2 API and the mental model for it as a GPU force-simulation engine, not a generic graph library. Auto-loads when working with cosmos.gl. Trigger on: "cosmos.gl", "cosmosgl", "@cosmos.gl/graph", "force graph", "WebGL graph", "setPointPositions", "setLinks", "setConfig", "setConfi
cosmos-mosaic-duckdbData layer for cosmos.gl: DuckDB-WASM ingestion, Mosaic Coordinator and Selections, MosaicClient subclassing, vgplot grammar, and the binding model that wires vgplot charts to the cosmos.gl graph through shared Selections. Trigger on: "DuckDB", "DuckDB-WASM", "Mosaic", "@uwdata/mosaic-core", "Coordi
cosmos-performanceHow to keep cosmos.gl smooth at 50K+ points: Float32Array reuse, batched setter calls, label budgeting, simulation warmup, GPU memory budget, capability detection (iOS EXT_float_blend, Android OES_texture_float), and instance lifecycle. Most cosmos.gl bugs are performance bugs dressed as visual bugs
cosmos-recipesBank of runnable annotated cosmos.gl + Mosaic recipes for archetypal design questions. Each recipe encodes a tested configuration with tuning notes, failure modes, and explicit "when to use" / "when NOT to use" sections. Trigger on: "cosmos.gl recipe", "force graph example", "pinned positions", "clu
cosmos-scene-directiveTranslation layer between Theseus's TF.js-driven scene intelligence and cosmos.gl's config surface. Owns the SceneDirective contract, the applyDirective adapter, the construction phasing (Galaxy, Filter, Build, Crystallize, Explore), and the pending-state visual for missing layer data. Trigger on: "
/cosmos-pro:cosmosCosmos-Pro hub: routes to the right cosmos.gl + Mosaic specialist (architect, data, render, chart, critic) and surfaces the right skill (foundations, mosaic-duckdb, recipes, scene-directive, performance) for the task at hand.
/cosmos-pro:learnRun the compound learning pipeline for cosmos-pro. Save the session log, run Bayesian confidence updates, and present the review queue with auto-captured claims about cosmos.gl, Mosaic, vgplot, SceneDirective adapter behavior, performance pitfalls, and color/token discipline. - Read - Edit - Write -
cosmos-architectPlans a cosmos.gl scene from a problem description. Default entry point for any cosmos.gl design task. Refuses to plan until three questions are answered: question shape, upstream data availability, view scope. Produces a structured plan covering position/weight/edge layers, vgplot charts, Mosaic Se
cosmos-chartAuthors a single vgplot chart (histogram, timeline, strip plot, dot plot, heatmap) bound to a named Mosaic Selection. Takes the selection name, data column, and intended interaction (brush, click, hover) and produces the vgplot spec. Simple charts are 20-line answers; complex ones (two-axis timeline
cosmos-criticReviews cosmos.gl + Mosaic + vgplot work after implementation. Runs every VERIFY check across the five cosmos-pro skills (foundations, mosaic-duckdb, recipes, scene-directive, performance). Flags performance smells, hardcoded colors, recipe drift (new code that duplicates an existing recipe's intent
cosmos-dataWires the data layer for a cosmos.gl scene. Given a plan from cosmos-architect, writes the DuckDB-WASM loading code, the Mosaic Coordinator setup, the table schemas, and the Selection graph between charts. Outputs TypeScript and SQL, not prose. Reads the actual Theseus Django API schema (via theseus
cosmos-renderWrites the cosmos.gl rendering layer: the canvas component, the CosmosGraphClient that bridges Mosaic and cosmos.gl, and the applyDirective adapter that translates Theseus SceneDirectives into cosmos.gl setter calls. Consumes a plan from cosmos-architect and the data layer from cosmos-data. Output i
Theorem's Harness: GraphQL-first MCP for adaptive planning, context, coordination, execution, validation, memory, and learning.
37 skills · 11 commands · 10 agents · 1 MCP server · hooks
JavaScript expertise plugin with 27 specialist agents covering frontend frameworks, D3.js visualization (force tuning, hierarchy, annotations, clustering, interaction, labels, style presets), Three.js / React Three Fiber / Drei 3D development, and access to real framework source code. Includes skills for JS design patterns, D3 standards, Three.js design, component APIs, and framework migration.
5 skills · 1 command · 27 agents
Intelligence engineering plugin for building systems that learn from their own operation. Successor to SciPy-Pro v4. Provides 7 workflow commands (/reason, /graph, /train, /architect, /simulate, /measure, /learn), 24 domain-expert agents across three tiers, and compound-learning + enrichment-pipeline guidance for EpiGNN, KGE RotatE, SBERT, GL-Fusion, model-swarm routing, and network effects.
1 skill · 7 commands · 24 agents
Animation & motion expertise across four domains: UI motion, creative/generative, 3D animation, and programmatic video. Backed by 21 reference source repos and specialized agents grounded in real implementations.
4 skills · 9 commands · 10 agents
Native iOS and macOS development plugin with deep expertise in SwiftUI, Swift 6 concurrency, SwiftData, XcodeBuildMCP integration, and Apple platform conventions.
12 commands · 9 agents · hooks
Django backend mastery: model architecture, ORM optimization, API design (DRF + Ninja), django-polymorphic patterns, MCP server construction, scientific Python integration, and Pydantic v2 mapping. Source-code-backed with Django ORM internals, DRF, Ninja, django-polymorphic, django-mcp-server, django-model-utils, django-pandas, django-filter, Pydantic v2, PydanticAI, and dj-notebook.
1 skill · 9 commands · 7 agents
'Expert in ALL computational collage composition: photo mosaics, grid layouts, scrapbook/journal styles, magazine editorial, vision boards, mood boards, social media collages, memory walls, abstract/generative arrangements, and art-historical techniques (Hockney joiners, Dadaist photomontage, Surrealist assemblage, Rauschenberg combines). Masters edge-based assembly, Poisson blending, optimal transport color harmonization, and aesthetic optimization. Activate on ''collage'', ''photo mosaic'', ''grid layout'', ''scrapbook'', ''vision board'', ''mood board'', ''photo wall'', ''magazine layout'', ''Hockney'', ''joiner'', ''photomontage''. NOT for simple image editing (use native-app-designer), generating new images (use Stability AI), single photo enhancement (use photo-composition-critic), or basic image similarity search (use clip-aware-embeddings).'
1 skill
This skill should be used when working with GDAL command line tools for raster and vector geospatial processing. Covers inspection, reprojection, clipping, raster conversion, mosaics, tiling, rasterization, and COG creation.
1 skill
KOZMOS Mosaic: a GitHub-style activity calendar across sessions. 26 weeks of days colored by turns, tool calls, tokens or dollars, with the current and longest streak, the best day, totals and today's line. /mosaic toggles the sidebar.
A mosaic that gains one tile per tool call, coloured by tool, until the session is a picture.
Databricks Free Edition — one $0, serverless-only, no-cloud end-to-end tutorial walked as 14 ordered stages: setup (CLI · UC · dbt) → lakebase (serverless Postgres 17 + Data API) → lakehouse (medallion 3 ways (notebook/dbt/SDP) + CDC) → analytics (Metric Views + AI/BI + Genie + ai_query) → retrieval (Mosaic AI Vector Search + pgvector-in-Lakebase over the contract docs). Start with /silver-databricks-end-to-end:start.
2 commands
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.