Kata Agent Team

Autonomous coding agents that continuously improve

An autonomous agent team that keeps getting better — organized as a daily Plan-Do-Study-Act cycle. Six agents. Sixteen skills. Zero infrastructure.

Scroll

Agent teams fail when they get complicated.

Most agent setups drown in infrastructure, sprawling toolchains, and unauditable prompt chains. Kata takes the opposite path — radical simplicity. Everything you need, nothing you don't.

16
Skills
Each under 200 lines of text
0
Infrastructure
No databases, no queues, no servers
0
Dependencies
Plain JavaScript, no third-party packages

One cycle. Every day.

Every workflow maps to a phase. Findings from Study always re-enter the loop — nothing is observed without downstream action.

P
Plan

Turn approved specs into architectural designs, then executable plans with steps, files, sequencing, and risks.

D
Do

Execute plans via implementation PRs. Run scheduled workflows that harden, release, and maintain. Every run captures a trace.

S
Study

Analyze outputs across four streams: security audits, feedback triage, documentation review, and grounded theory from traces.

A
Act

Trivial findings become fix PRs. Structural findings become spec documents. Fix and spec branches never mix.

Six agents. Explicit scope.

Each persona knows what it must do — and what it must not. When a finding exceeds scope, the agent writes a spec rather than attempting the fix.

📐
Staff Engineer
Plan · Do

Owns the full spec, design, plan, implement arc for approved specs.

📦
Release Engineer
Do

Keeps PR branches merge-ready, repairs CI, cuts releases. The sole external merge point.

🔒
Security Engineer
Do · Study · Act

Patches dependencies, hardens supply chain, enforces security policies.

📋
Product Manager
Study · Act

Triages issues against product vision, reviews spec quality, runs evaluations.

📝
Technical Writer
Study · Act

Reviews docs for accuracy, curates agent memory, fixes staleness.

Improvement Coach
Study

Facilitates the daily storyboard meeting and 1-on-1 coaching sessions.

Same agents. Every surface.

The same profiles and skills operate identically whether triggered by your IDE, a cron schedule, a GitHub event, or a bridged message.

💻
IDE
Direct invocation
Scheduled Shifts
Cron → kata-shift
💡
GitHub Issues
Event → kata-dispatch
🔀
GitHub PRs
Event → kata-dispatch
💬
GitHub Discussions
Bridge → kata-dispatch
📨
Microsoft Teams
Bridge → kata-dispatch

One Git repo of markdown files.

Every agent reads and writes the same wiki — priorities, logs, metrics, and storyboards. A scheduled shift, a bridge-dispatched message, and an IDE session all share the same state. No database. Just markdown in a Git repository.

Three lines. That's it.

Install the skill pack. Tell Claude to set up the team.

Terminal
cd my-repo/
apm install forwardimpact/kata-skills
echo "Setup the Kata Team" | claude