prototype-openclaw-tui
Skill59 starsChanged 16 days agoMITView on GitHub
Build throwaway, fixture-driven OpenClaw Clack or Pi TUI prototypes and compare multiple interactive variants side by side in tmux without running the full application or touching live state.
We split this skill into the rules it gives an agent
17 rules. About 0 are common, 12 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈12
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 10
Use this skill to isolate one OpenClaw terminal surface, render it with the real OpenClaw TUI stack, and compare two to six design variants …
- Distinctive: line 15
Read the root and scoped AGENTS.md files before editing.
- Very specific: line 16
Keep the prototype beside the target and name it *.prototype.ts.
- Very specific: line 17
When iterating on an existing surface, include its current implementation as the first baseline variant with the same fixture data.
- Distinctive: line 20
Use the real OpenClaw renderer, theme, copy, and component where practical.
- Distinctive: line 23
Do not run the full OpenClaw application when the isolated surface is enough.
- Distinctive: line 24
Treat prototype code as throwaway.
- Distinctive: line 34
Do not invent a shared Clack/Pi abstraction.
- Distinctive: line 39
Read the target module, its renderer adapter, callers, and adjacent tests.
- Distinctive: line 40
Create the smallest executable harness that reaches the target surface.
- Very specific: line 41
If the surface already exists, make the no-argument invocation render its current implementation unchanged as --variant=baseline.
- Very specific: line 43
Add structurally different alternatives behind --variant=<id>, for two to six total variants including the baseline.
- Distinctive: line 46
Run each variant directly in its own tmux pane.
- Distinctive: line 51
After the user chooses a direction, carry the decision into production code with normal tests and validation.
- Distinctive: line 56
Run the bundled launcher with a session name, repository path, then title and command pairs:
- Very specific: line 68
After editing the prototype, run the same command with --refresh instead of --open.
- Distinctive: line 94
Leave the session running for user review unless asked to stop it.
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Prototype OpenClaw TUI
Use this skill to isolate one OpenClaw terminal surface, render it with the real OpenClaw TUI stack, and compare two to six design variants in tmux.
Rules
- Read the root and scoped
AGENTS.mdfiles before editing. - Keep the prototype beside the target and name it
*.prototype.ts. - When iterating on an existing surface, include its current implementation as
the first
baselinevariant with the same fixture data. Only omit the baseline when the user explicitly asks for a greenfield comparison. - Use the real OpenClaw renderer, theme, copy, and component where practical.
- Replace scanning, models, network calls, installs, config writes, SQLite, and other durable effects with deterministic in-memory fixtures.
- Do not run the full OpenClaw application when the isolated surface is enough.
- Treat prototype code as throwaway. Do not promote it directly to production.
- Preserve unrelated and pre-existing worktree changes.
Choose the renderer
- Wizard, onboarding, configuration, doctor, selection, or progress surface:
use
createClackPrompter()fromsrc/wizard/clack-prompter.ts. - Agent shell, chat, overlay, editor, selector, or
src/tui/**component: use@earendil-works/pi-tuiand the existing OpenClaw component.
Do not invent a shared Clack/Pi abstraction. They are separate prototype recipes joined only by the tmux comparison loop.
Workflow
- Read the target module, its renderer adapter, callers, and adjacent tests.
- Create the smallest executable harness that reaches the target surface.
- If the surface already exists, make the no-argument invocation render its
current implementation unchanged as
--variant=baseline. - Add structurally different alternatives behind
--variant=<id>, for two to six total variants including the baseline. - Give every variant the same fixture data and terminal dimensions.
- Run each variant directly in its own tmux pane. Do not pipe interactive TUI output; tmux must provide the PTY.
- Iterate on the harness with the launcher's
--refreshmode. Respawn the panes in the existing session; do not kill the session or reopen the user's terminal. - After the user chooses a direction, carry the decision into production code with normal tests and validation. Remove or separately capture the prototype.
Launch the comparison grid
Run the bundled launcher with a session name, repository path, then title and command pairs:
.agents/skills/prototype-openclaw-tui/scripts/launch-tmux-grid.sh \
--open \
app-recommendations-prototype "$PWD" \
"A - Baseline" "node --import tsx src/wizard/setup.app-recommendations.prototype.ts --variant=baseline" \
"B - Grouped" "node --import tsx src/wizard/setup.app-recommendations.prototype.ts --variant=grouped" \
"C - Focused" "node --import tsx src/wizard/setup.app-recommendations.prototype.ts --variant=focused"After editing the prototype, run the same command with --refresh instead of
--open. This preserves the tmux session, attached external terminal, window,
and first pane while restarting all pane commands and reapplying the grid:
.agents/skills/prototype-openclaw-tui/scripts/launch-tmux-grid.sh \
--refresh \
app-recommendations-prototype "$PWD" \
"A - Baseline" "node --import tsx src/wizard/setup.app-recommendations.prototype.ts --variant=baseline" \
"B - Grouped" "node --import tsx src/wizard/setup.app-recommendations.prototype.ts --variant=grouped" \
"C - Focused" "node --import tsx src/wizard/setup.app-recommendations.prototype.ts --variant=focused"The launcher refuses to replace an existing session unless --refresh is
explicitly provided. --open uses the user's
.command-associated terminal app on macOS, Windows Terminal from WSL, or the
first available Linux launcher from xdg-terminal-exec, $TERMINAL, and
x-terminal-emulator. If no supported launcher is available, the tmux session
remains ready at equal detached dimensions and the script exits successfully
after printing the manual attach command. Attach, then run --refresh once to
adopt the terminal's larger dimensions. The same guidance applies if an
external terminal opens but does not attach within ten seconds. When invoked
from a non-interactive TERM=dumb environment, the launcher removes its
inherited NO_COLOR only from prototype pane processes so the external
terminal can detect and render its normal colors. Prefer the external terminal
over the Codex in-app terminal unless the user asks for the in-app surface.
Leave the session running for user review unless asked to stop it.
Handoff
Report:
- the prototype path and exact run command;
- the tmux session name and attach command;
- which effects were replaced with fixtures;
- available variant ids;
- whether the prototype changed production code (normally no).
Other versions10 projects use a version of this.
Word for word · 10
- steipete/warelay★ 394k100% same
- moltbot/clawdbot★ 394k100% same
- moltbot/moltbot★ 394k100% same
- clawdbot/clawdbot★ 394k100% same
- steipete/clawdis★ 394k100% same
- openclaw/openclaw★ 394k100% same
- molt-bot/clawdbot★ 391k100% same
- friuns2/openclaw-android-assistant★ 573100% same
- OpenClawAndroid/openclaw-android-assistant★ 556100% same
- supermansuper563-ops/openclaw-workbench★ 0100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
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 repository107 other files it gives its agents.
- AGENTS.mdAGENTS.md
- docs/AGENTS.mdAGENTS.md
- extensions/AGENTS.mdAGENTS.md
- scripts/AGENTS.mdAGENTS.md
- test/AGENTS.mdAGENTS.md
- ui/AGENTS.mdAGENTS.md
- agent-transcriptSkill
- auto-qaSkill
- autoreviewSkill
- channel-message-flowsSkill
- clawdtributorSkill
- claw-scoreSkill
- clawsweeperSkill
- control-ui-e2eSkill
- crabboxSkill
- discord-clawdSkill
- discord-e2eSkill
- discord-user-postSkill
- discrawlSkill
- gitcrawlSkill
- graincrawlSkill
- notcrawlSkill
- openclaw-changelog-updateSkill
- openclaw-ci-limitsSkill
- openclaw-debuggingSkill
- openclaw-docker-e2e-authoringSkill
- openclaw-ghsa-maintainerSkill
- openclaw-live-updaterSkill
- openclaw-parallels-smokeSkill
- openclaw-pr-maintainerSkill
- openclaw-qa-testingSkill
- openclaw-refactor-docsSkill
- openclaw-release-validationSkill
- openclaw-repair-sweepSkill
- openclaw-secret-scanning-maintainerSkill
- openclaw-test-heap-leaksSkill
- openclaw-testingSkill
- openclaw-test-performanceSkill
- openclaw-updateSkill
- parallels-discord-roundtripSkill
- plugin-sdk-impactSkill
- proof-videoSkill
- pr-simplifySkill
- release-openclaw-announcementSkill
- release-openclaw-ciSkill
- release-openclaw-macSkill
- release-openclaw-maintainerSkill
- release-openclaw-plugin-testingSkill
- security-triageSkill
- slack-e2eSkill
- slacrawlSkill
- solidSkill
- tag-duplicate-prs-issuesSkill
- technical-documentationSkill
- telegram-e2e-userbotSkill
- test-auditSkill
- update-team-serverSkill
- verify-releaseSkill
- 1passwordSkill
- apple-notesSkill
- The first 60 shown.
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.

