whiteboard
Skill1 starsChanged 57 days agoMITView on GitHub
A relentless interview to sharpen a plan or design, held over a live Excalidraw canvas you can draw on, which also creates docs (ADRs and glossary) as we go.
6 other projects use a version of this.
Adopted by
Most recently taken up by
- nearmap/shipyardtheir own version, first seen Nov 8, 2026
- martincjarvis/development-skillstheir own version, first seen Nov 8, 2026
- edenfunf/reelmimictheir own version, first seen Nov 4, 2026
- Johnwang688/jarvistheir own version, first seen Oct 13, 2026
- markphelps/agent-pluginstheir own version, first seen Sep 28, 2026
- FinnovateIo/skillstheir own version, first seen Aug 5, 2026
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- edenfunf/reelmimic★ 1.8ksame name
- FinnovateIo/skills★ 8same name
- markphelps/agent-plugins★ 5same name
- martincjarvis/development-skills★ 4same name
- nearmap/shipyard★ 1same name
- Johnwang688/jarvis★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository27 other files it gives its agents.
- blast-radiusSkill
- codebase-designSkill
- code-reviewSkill
- create-verification-skillSkill
- designingSkill
- domain-modelingSkill
- handoffSkill
- hf-cliSkill
- howSkill
- implementSkill
- improve-codebase-architectureSkill
- interrogateSkill
- laziness-protocolSkill
- maintain-verification-skillSkill
- navigatorSkill
- no-commentsSkill
- prototypeSkill
- recallSkill
- redesign-from-first-principlesSkill
- researchSkill
- review-unpushedSkill
- scopeSkill
- show-me-your-workSkill
- technical-writingSkill
- tracerSkill
- unslopSkill
- writing-for-agentsSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 20
Call the Skill tool for designing and domain-modeling at the start.
- Distinctive: line 50
When a decision settles, recolour it.
“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Whiteboard
An interview you can draw on. The canvas is where the design tree becomes visible; the durable output is still ADRs and a glossary.
Three skills, one session:
| owns | |
whiteboard | the canvas: what is drawn, and reading back what the user drew |
designing | the interview: the design tree, the frontier, the rounds |
domain-modeling | the record: ADRs and CONTEXT.md |
Call the Skill tool for designing and domain-modeling at the start. They run
unchanged. This skill adds a surface to them, and takes no decisions of its own.
Read canvas/PROTOCOL.md before drawing anything. It
holds the rules for sharing a canvas with someone who is drawing on it too: read
before you write, red is theirs, back up first, recolour rather than redraw. They
are not optional and they are not repeated here.
The canvas is optional. Where the whiteboard MCP server is not configured or
its container is down, say so in one line and run the interview text-only. That is
what this skill was before it could draw, and nothing below is load-bearing for it.
What the colours mean
| state | colour | meaning |
| settled | #2f9e44 green | decided, and written into an ADR |
| open | #1971c2 blue | on the frontier, asked this round |
| blocked | #868e96 grey | waiting on a decision above it |
| theirs | #e03131 red | never write it |
Drawing the design tree
designing produces a tree of decisions worked in rounds, where the frontier is
every decision whose prerequisites are settled. That tree is the thing to draw.
- One node per decision. Number it to match the
Q1,Q2of the round, so an answer in chat and an answer on the canvas mean the same thing. - An edge from each decision to the decisions that hang off it.
- Colour by state.
- When a decision settles, recolour it. Do not rebuild the tree.
Draw the round's questions, then ask them in chat as designing specifies. The
canvas does not replace the written round; it gives the user somewhere to answer
that is faster than prose.
Reading their answers
Everything in the protocol's four-signal table applies, and here is what each one settles:
| signal | what it means for the tree |
| red note on a node | the answer, or a constraint on it — restate it and confirm |
| a box they drew | a decision the tree was missing; add it and re-derive the frontier |
| a node moved beside another | the two are one decision, or the order is wrong |
| a node deleted | the question does not apply. Do not re-ask it |
An answer given on the canvas is an answer. Carry it into the round as though they had typed it.
Ending the session
The canvas is scratch, so it ends as a file rather than as a live service:
export_scene filePath: docs/app-maps/efforts/<slug>.excalidrawThat path sits beside the effort doc navigator keeps, so a later session picks
up both together.
Then the real outputs, through domain-modeling: an ADR per decision settled,
and the terms the session argued about in CONTEXT.md. A session that ends with
only a drawing has produced nothing.
Where this goes next
whiteboard is the unmapped conversation. When the effort is bigger than one
session and the project keeps an APP_MAP.json, navigator charts it as decision
tickets and draws the route graph on this same canvas — see
../navigator/references/route-canvas.md.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.