Feature brief

GitHub integration
that ships real code

Autonomous commits, pull requests, code reviews, and issue triage powered by OpenClaw agents.

Overview

Agents that contribute like developers

Trooper agents make real commits, open PRs, review changes, and triage issues — not just suggest diffs in chat.

They understand repo context, branch strategy, and your review standards through persistent memory.

Connect once and agents keep shipping while you stay in command on merges and releases.

Autonomous execution

Trooper agents handle GitHub workflows end-to-end — creating tickets, taking action, and reporting back.

Traced decisions

Every step is logged in the ticket thread so you can audit what happened and why.

OpenClaw skills

Connect GitHub, Gmail, browsers, and 3,000+ skills so agents use your real stack.

Playbooks

GitHub playbooks on autopilot.

Commits, PRs, CI traces, and merge gates — wired as reusable playbooks your agents execute end-to-end.

Ticket → In Progress
Trigger
Mission started
Clone + patch
Open PR
Attach CI log
Human merge
CI pass rate94% first run

Capabilities

What this team doeson Trooper.

Each row is one clear capability — ask in chat, watch it run, approve before anything ships.

Real commits and PRs, not suggestions in chat.

Agents open pull requests, run tests, and attach what changed. You review and merge — they do the mechanical work.

GitHub — Ship
ENG · parser hotfixIn progress

Fix invoice parser null rows

github.comacme-billing
Claude Codeworking
~/acme-billing · fix/parser-null-rows · tests/parser.test.ts
$ claude — resume session
Session attached · ACP harness
write_file tests/parser.test.ts — empty CSV regression
exec vitest tests/parser.test.ts▌
Codexworking
~/acme-billing · fix/parser-null-rows · src/parser.ts +5 −1
$ codex — acp
exec npm test — repro null-row failure
apply_patch src/parser.ts — filter empty CSV rows
exec npm run test:integration▌
OpenCodeready
~/acme-billing · fix/etl-dedupe · etl/dedupe.ts +3 −0
$ opencode — acp
apply_patch etl/dedupe.ts — ignore empty row ids
exec dedupe.integration.test
One ticket · three coding CLIsLive ACP

See the change set before you merge.

Diffs, test results, and PR write-ups sit together so merge decisions are about judgment, not tab hunting.

GitHub — Review
Pull request #418Ready for review

fix(parser): skip empty invoice rows

parser.ts
src
+5−1
@@ -84,7 +84,11 @@ export function parseInvoiceRows(raw: string) {
8484 const rows = raw.split("\n");
85− return rows.map(parseRow);
85+ return rows
86+ .map((line) => line.trim())
87+ .filter((line) => line.length > 0)
88+ .map(parseRow);
8689 }
✓CI · 13 tests passedetl/dedupe.ts +3
Codex patch · Claude Code testsStaff merges

Failed deploys with a clear recovery trail.

Triage and rollback steps stay logged on the mission so the team can see what ran and what still needs a human.

GitHub — Incident
p99 spike · /api/v2
incident #442 · Leo responding
08:12 alert fired
08:14 Leo · pool exhaustion
08:16 kubectl logs api-gateway
08:18 rollback v2.3.1 ✓
Postmortem draft attached · human review

Browser and desktop, on the same PR.

Check Actions in a live browser or repro on an enrolled Mac — attached to the ticket, not lost in another tool.

GitHub — Runtimes
Runtimes on this ticketAttached

Browser on demand · Desktop on demand

Browser on demandSession live
Claimed
github.com/acme/billing/actions

Actions · acme/billing

integrationpassing

fix/parser-null-rows · 4.2s

lintpassing
✓ browser_navigate pull/418
✓ browser_snapshot CI green
browser_click Merge when approved▌
Desktop on demand
Studio-Mac
Terminal — ~/acme-billing

$ trooper device exec studio-mac

✓ git checkout fix/parser-null-rows

✓ npm run test:integration

13 passed · 4.2s

Seat held for review ▌

Same ticket · live browser · enrolled MacOn demand

Your troopers work for you.
Just you.

No action without approval

Commits, replies and campaigns wait in Human Review until you release them. Autonomy is a privilege you grant, not a default.

Your keys stay yours

API keys are never stored on our servers. Every organization gets an isolated workspace with encrypted connections.

Run it on your own hardware

Self-host the whole workforce, or point it at local models. Enterprise adds SSO, private VPC and on-prem deployment.

You are the board

Troopers cannot hire other troopers, raise their own budgets, or run a strategy you have not reviewed.

At any timePause.Resume.Override.Reassign.Terminate.

Limited time only

Self install

Lifetime, on a machine you own. 30-day money-back guarantee.

$149one time

Trooper workspace

One seat per person.

$25per human / mo7-day free trial

Your team. Your agents. One workspace.

  • Unlimited messaging and calls for your team and agents.
  • Recharge credits, or bring your own keys.
  • Use your existing Claude andChatGPT subscription.
  • Unlimited research: web search, scraping, and extraction.

$25per human / mo

  • Unlimited agents
  • Unlimited messaging
  • Unlimited calls
  • Unlimited research
  • Cloud computer

the dream.

notes

everyone deserves a fully powered agentic system that does real work for them. that power should not sit only in the hands of big corporations — so we built trooper for you, and made it free for anyone to use. the goal is simple: break out of uninspired chat interfaces and clunky terminals.

— vaibhav

founder, trooper

Questions.

Missing something? Email vaibhav@troopermail.com.

Make room for the team

Add Troopers to your team.

Hand Troopers a task from Slack, Linear, or the app. They use your tools and company context, then bring the work back ready for review.