MisterGC

Compose Clayground labs — deterministic, agent-verifiable teaching and experiment spaces — from the clay_lab kernel and the domain kits, and author the full lab triad (Sandbox.qml + paper.md + overview.grafli + strings.js + flows). Use this skill when creating or extending anything under labs/ (labs or kits), when working with Clayground.Lab components (Lab, Parameter, Probe, SimClock, Plot2D, Flow, Narrator, LabTheme, LabLang), or when the user asks for a new lab, kit, teaching space, or simulation playground. Companion of clay-crew, which owns verification through the inspector — load clay-crew whenever you verify.

Collaborate on and verify Clayground QML sandboxes through the Dojo's built-in inspector. Use this skill when working on Clayground QML applications and needing to verify results, diagnose load failures or crashes, read user flag feedback, search for entities, trace behavior, or collaborate on visual output. Triggers when the context involves .clay/inspect, .clay/crew, flagInfo, ClayInspector, canvas.find, trace, or the user mentions verifying sandbox output.

Author and edit `.grafli` files — plain-text, line-oriented diagrams rendered by the grafli desktop app. Trigger on visualization requests, explicit or indirect: "draw / diagram / sketch / visualize / map out / graph this", "show me as a diagram", "make a grafli", "put it on a board / whiteboard", or when working on existing `.grafli` files. ALSO trigger when asked to EXPLAIN, walk through, narrate, tour, or PRESENT an existing diagram/graph (author `@ bookmark` / `@ flow` guided tours, exportable as slides), and on THINKING-BOARD requests — "help me think through X", "weigh the options", "map the unknowns", "should we do A or B" — where a human would go to a whiteboard: author a board that structures the problem. Do NOT trigger on generic "review this code", "explain this function", "summarize this module" requests unless the user also asks for a visual, a board, or a diagram. When unsure, ask the user a one-line clarifier before pulling this skill in.

Author and review Markdown documents that a human reads, annotates, and navigates in textli — a distraction-free editor whose reading view renders the file with typeset pandoc math, followable `path:line` source references, and inline CriticMarkup review (comments and track-change suggestions). Trigger when handing a Markdown document to a human for reading or review (design docs, reports, papers, "write it up so I can review it"), when a Markdown file contains CriticMarkup marks ({>>…<<}, {++…++}, {--…--}, {~~…~>…~~}) to process, when asked to review or annotate a human's draft, or when writing a document genre covered by references/ (decision doc, learning doc, scientific paper). Do NOT trigger for ordinary Markdown nobody will process in textli (READMEs, changelogs, code comments).