Squads are teams of AI agents, defined in markdown and tracked in git. They pick up work from your issues and open pull requests for your team to review.
You set the outcome. Squads plan and do the work, checks run on every result, and what reaches you is the part that needs a decision.
Tell a squad what you need in plain language. No prompts to engineer, no workflows to wire up.
"Find 20 mining suppliers hiring in Chile this quarter."
Specialized agents pick up the work and use real tools — web, databases, GitHub, your CLI. Each step is logged.
search → extract → validate → draft
Results land as a pull request, a report or a dataset, with their checks already run. You decide what ships.
CSV + sources, ready to act on.
Each run saves notes to plain files in git. The next run can read them, and you can read, edit, or revert them.
next run: starts from the saved notes.
Agents pick up GitHub issues, write code, open pull requests, and report back. Each pull request waits for your review.
Agents read an issue, plan a solution, and open a pull request for you to review.
Issues become tasks. Pull requests become deliverables. Your existing workflow stays intact.
Agents propose. You approve. Every action is reviewable before it goes live.
Each agent's mission, boundaries, and approval gates are written in a markdown file you can read and edit. Every run is logged, so you can check what it did.
Every squad has a clear mission. Every agent has a defined role. Work flows from goals, not ad-hoc prompts.
Agents run in an OS sandbox by default. Credential stores are denied and destructive git commands are blocked.
Decisions, mistakes, and insights are saved as plain files that later sessions can read. You can read, edit, or delete them.
Tell us about your work. We'll go through your workflows with you and identify which are specific and checkable enough to hand to agents.
30-minute call — no commitment