Skip to content
Agentmart
Submit a tool

waterfall

Implement anything end-to-end using two agents: planner cum reviewer (premium model) and an implementer (fast model)

Skill by juspay

IndexedMIT★ 11
Write a review

Chalk Labs

Need help deploying waterfall?

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.

Book a call

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/juspay-waterfall

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.

We use your email only for this newsletter. Privacy

About

Implement anything end-to-end using two agents: planner cum reviewer (premium model) and an implementer (fast model)

No 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/waterfall

What's inside

Skill name
waterfall
Description
Implement anything end-to-end using two agents: planner cum reviewer (premium model) and an implementer (fast model)
License
MIT
Compatibility
Requires kolu
Instructions
27 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Waterfall

This skill is invoked with an argument that tells us what to implement.

You will respond to the user with a plan (focusing more on architecture-level than walls of text), who then approves it (unless they pre-approve). After they approve you must finalize the plan to remove all ambiguities using the question tool, along with asking the user these questions:

- Which harness to run (e.g.: omp, codex --yolo, etc.)
- Which model to use for the selected harness
- Whether to auto-merge on green CI at end

Then you open a new Kolu split terminal, and run the implementor agent (in the same $PWD worktree as you) to implement the plan opening PR[^pr] and then wait[^debrief] for the implementor to finish. Then, you review the PR per repo's guidelines (e.g.: [Cordis-perfection](https://github.com/juspay/olai/blob/master/docs/architecture/cordis.md), [/solidjs](https://github.com/juspay/skills/blob/main/skills/solidjs/SKILL.md)) posting it in the PR itself, and then ask the implementor to address it. Then you re-review. Once satisfied, you ask the implementor do a final refactor per https://kolu.dev/blog/hickey-lowy/ and then to run CI (to save time, we don't run full CI until reviews are fully done). Once the PR is green, our work is done (unless auto-merge is enabled).

**IMPORTANT**: 
- If you are Fable model, you must **NOT** use Fable for any subagents you spawn unless that subagent requires premium intelligence.
- Keep it cheap: the brief holds only what this skill and the repo rules require; tests run once per round and CI once at the end; mid-round asks wait for the next round; compact the implementor only when the context is large, or before a long new phase like CI

**BANS**:
- You are banned from doing any implementation yourself; that happens only through the implementor agent.
- You are banned from starting new agents or new PRs without human's approval.
- You are banned from terminating the implementor agent terminal *unless* the PR gets merged or closed.

[^debrief]: Use kolu debrief to wait on any kolu terminal running an agent to go idle.
[^pr]: PR ttle & body should be interesting to read to users, not boring.

From juspay/skills, MIT.

More from juspay/skills

Signals

Stars
11
Forks
0
Open issues
3
Last commit (this folder)
Oct 13, 2026 (19 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from juspay

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 waterfall 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.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying waterfall?

CallGet help
Cite this page

waterfall skill. Agent Mart. https://agentmart.dev/skills/juspay-waterfall (last updated 2026-10-24)