prototype
Submit a component design proposal to cueloop for human review without blocking.
Skill by mmurakaru
Chalk Labs
Need help deploying prototype?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/mmurakaru-prototype
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: design, frontend, backend
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/prototype
What's inside
- Skill name
prototype- Description
- Submit a component design proposal to cueloop for human review without blocking. Use when the user asks to review a UI component's shape before implementing it: its prop/type API, composition, and callstack.
- License
- Apache-2.0
- Instructions
- 94 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# cueloop prototype review
Review a UI component on paper before you build it: its interface, how it is
assembled from primitives you already have, and its callstack. You write a single
Markdown design doc; the reviewer reads it in the terminal and annotates the exact
lines. The review does not block: you submit, keep working with the user, and
cueloop delivers the Message when the reviewer is done.
Write the doc as plain Markdown with exactly these three sections, in order. Keep
everything line-addressable text, since the reviewer anchors comments to lines and
words. Prefer fenced code blocks and indented trees over prose. Do not use Markdown
tables; the reviewer renders one as a single block, so write comparisons as a list.
## 1. API
The prop/type surface, as a fenced `ts` block: the real interface a caller sees.
Extract it from the component's types, or write the proposed one. Include every
prop, its type, whether it is optional, and a one-line intent where it is not
obvious.
## 2. Composition
How the component is assembled, as a **JSX anatomy tree** in a fenced `tsx` block,
following the react-aria / Radix "Anatomy" convention. Write the real element tree,
nesting by indentation, and put a trailing `{/* ... */}` comment on a node to name
the primitive it reuses (or mark it a slot). It reads as code, so the reviewer
annotates actual JSX lines. When there is more than one way to build it, list the
options and your recommendation as a list, not a table:
```tsx
<ComboBox> {/* generic over T */}
<Label />
<Group>
<Input /> {/* reuse: Input */}
<Button /> {/* toggle the popover */}
</Group>
<Popover> {/* reuse: Popover, focus-trap + esc */}
<ListBox>
<ListBoxItem /> {/* active / selected */}
</ListBox>
</Popover>
</ComboBox>
```
Options considered:
- **A, reuse the shared Popover**: focus-trap and escape-to-close come for free; least new code.
- **B, a bespoke popover**: full control of placement, but it re-solves focus management.
Recommendation: A. Reuse the Popover and keep the ComboBox to input + listbox wiring.
## 3. Callstack
The call and data flow through the component, as an indented tree in a fenced block:
entry point at the top, each call nested one level deeper, down to the store or
output. Designing this flow up front is the point of the review. It shows how a
user action threads through the layers before any code exists.
```
main.ts
Command.runWith(rootCommand)
cli.ts alias command handler
parseLinkSlug("gh")
parseDestinationUrl("https://example.com/acme")
LinkManagerClient.alias({ slug, destination })
HttpApiClient client.operator.createAlias({ params, payload })
PUT {baseUrl}/api/links/gh
Access middleware
ProvideOperatorIdentity middleware
OperatorHandlers.createAlias
LinkCatalog.getLink(slug)
LinkCatalog.createAlias({ slug, destination, operator })
LinkCatalogStore.insertAlias
RedirectIndexService.putDestination
CliOutput.link(link)
```
## Submit and respond
Write the doc to one Markdown file and submit it through the harness's cueloop
`prototype` workflow when available. Otherwise, run
`cueloop session create --type prototype --content-file <design.md>` and give
the user the returned Thread ID. On the CLI path, collect the Message with
`cueloop session wait <id> --timeout-ms 60000`, repeating while pending.
On approval, proceed with the component. When changes are requested, apply
every edited line and annotation to the design doc, then revise the same
Thread with `cueloop session submit-revision <id> --content-file <design.md>`
on the CLI path.
Reviewing a rendered pixel mockup instead of a design doc is an opt-in
experimental mode, set by the `[experimental]` config section. It needs a
graphics-capable terminal and an installed browser.From mmurakaru/cueloop, Apache-2.0.
More from mmurakaru/cueloop
- 0.626score
annotate
mmurakaru
Read a cueloop plan or diff and comment on it as a review-side agent, without leaving the review.
SkillAFreshness: A★ 8Claude Code +9 - 0.626score
diff
mmurakaru
Submit working-tree changes to the cueloop changes panel.
SkillAFreshness: A★ 8Claude Code +9 - 0.626score
refine
mmurakaru
Analyze past cueloop Threads and propose recurring feedback writebacks through a plan Thread.
SkillAFreshness: A★ 8Claude Code +9 - 0.624score
review
mmurakaru
Review a GitHub pull request in cueloop, record agent findings on the diff, and refine them with the user before any explicit GitHub post.
SkillAFreshness: A★ 8Claude Code +9 - 0.622score
reply
mmurakaru
Submit the previous reply to a cueloop Thread for human review.
SkillAFreshness: A★ 8Claude Code +9 - 0.622score
plan
mmurakaru
Submit a Markdown plan to a cueloop Thread for human review.
SkillAFreshness: A★ 8Claude Code +9 - 0.620score
cueloop
mmurakaru
Review coding-agent plans, replies, diffs, prototypes, or pull requests in a terminal Thread and return human comments as a Message.
SkillAFreshness: A★ 8Claude Code +9 - 0.616score
pair
mmurakaru
Start a live cueloop pairing session where you work through the user's code comments.
SkillAFreshness: A★ 8Claude Code +9 - 0.602score
code-review
mmurakaru
Review the changes since a fixed point (commit, branch, tag, or merge-base) along two axes: Standards (does the code follow this repo's documented coding…
SkillAFreshness: A★ 8Claude Code +9
Signals
- Stars
- 8
- Forks
- 0
- Open issues
- 16
- Last commit (this folder)
- Sep 22, 2026 (1 month ago)
- Last release
- Oct 20, 2026 (16 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- Apache-2.0
- Repository
- github.com/mmurakaru/cueloop
- Website
- www.cueloop.dev
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from mmurakaru
- 0.626score
annotate
mmurakaru
Read a cueloop plan or diff and comment on it as a review-side agent, without leaving the review.
SkillAFreshness: A★ 8Claude Code +9 - 0.626score
diff
mmurakaru
Submit working-tree changes to the cueloop changes panel.
SkillAFreshness: A★ 8Claude Code +9 - 0.626score
refine
mmurakaru
Analyze past cueloop Threads and propose recurring feedback writebacks through a plan Thread.
SkillAFreshness: A★ 8Claude Code +9 - 0.624score
review
mmurakaru
Review a GitHub pull request in cueloop, record agent findings on the diff, and refine them with the user before any explicit GitHub post.
SkillAFreshness: A★ 8Claude Code +9 - 0.622score
reply
mmurakaru
Submit the previous reply to a cueloop Thread for human review.
SkillAFreshness: A★ 8Claude Code +9 - 0.622score
plan
mmurakaru
Submit a Markdown plan to a cueloop Thread for human review.
SkillAFreshness: A★ 8Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get prototype running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
prototype skill. Agent Mart. https://agentmart.dev/skills/mmurakaru-prototype (last updated 2026-10-24)