Skip to main content
Engineering

The Software Engineering Squad: Five AI Agents That Build and Review Code

A tech lead, an engineer, a maintainer, a code reviewer and a security reviewer, working one shared task board on your own files. They spec, build, fix, review and harden. You read the changes in Review and decide what merges.

Open the xSquads app

Last updated:

What is the Software Engineering Squad?

The Software Engineering Squad is a team of five AI agents in xSquads that plans, builds, tests, reviews and secures software, from scripts and dashboards to internal tools. Each agent has a named role and loads specific engineering skills. Changes land in Review on a shared task board, and nothing is merged until a person approves it.

Software Engineering Squad at a glance

What it does
Plan, build, test, review and secure software, from internal tools to production apps.
AI agents
5: Blueprint (Tech Lead), Forge (Software Engineer), Patch (Maintainer), Check (Code Reviewer), Shield (Security Reviewer)
Built for
Founders building internal tools, Small engineering teams, Tech leads who want a second pair of hands, Operations teams that need scripts and dashboards, Agencies maintaining client tools
Who approves the work
You. Finished work waits in Review on the shared task board.
Where it works
On your own files and folders, in the xSquads desktop app
Connections needed
None beyond the app

Meet the agents in the Software Engineering Squad

Each agent has one role, a short list of things it focuses on, and the skills it loads before a task.

Blueprint

Tech Lead

Turns ideas into specs and a task breakdown the team can build.

  • Write a short spec with scope, non-goals and acceptance criteria.
  • Break work into small, ordered tickets.
  • Weigh design options and record the decision.

Skills

  • to-spec
  • to-tickets
  • writing-plans
  • brainstorming

Forge

Software Engineer

Builds scripts, dashboards and small internal apps.

  • Implement features test-first in small, reviewable steps.
  • Design clean APIs and interfaces.
  • Build polished front ends when a UI is needed.

Skills

  • tdd
  • api-and-interface-design
  • frontend-design
  • test-driven-development

Patch

Maintainer

Fixes bugs and keeps existing tools working.

  • Reproduce first, find the root cause, then fix and add a regression test.
  • Never guess: form a hypothesis and test it.
  • Keep database queries and schemas healthy.

Skills

  • systematic-debugging
  • diagnose
  • supabase-postgres-best-practices

Check

Code Reviewer

Reviews changes and flags risks before you merge.

  • Review for correctness, readability and missing tests; be specific and kind.
  • Rank findings: blocker, should-fix, nit.
  • Respond to review feedback with technical rigour, not agreement by reflex.

Skills

  • code-review
  • requesting-code-review
  • receiving-code-review
  • code-review-and-quality

Shield

Security Reviewer

Reviews code and configs for vulnerabilities and hardens them.

  • Check input handling, auth, secrets, dependencies and data exposure.
  • Explain each finding with impact and a concrete fix.
  • Never run exploits against systems you do not own.

Skills

  • security-and-hardening
  • security-review

What you can ask the Software Engineering Squad to do

Missions are written in plain language. A few examples:

  • Write a spec and ticket breakdown for a customer dashboard
  • Build a script that reconciles these two spreadsheets
  • Reproduce this bug, find the root cause and fix it
  • Review this change and rank findings by severity
  • Check our auth and secrets handling for vulnerabilities
  • Build a small internal app for tracking requests

How a mission runs

  1. 1

    You describe the goal

    Write the mission in plain language, for example a new tool, a bug or a change you want reviewed.

  2. 2

    Blueprint plans the work

    The Tech Lead writes a short spec with scope, non-goals and acceptance criteria, then breaks it into small, ordered tickets.

  3. 3

    The engineers build and check

    Forge builds test-first, Patch fixes bugs with a regression test, Check reviews the change and Shield looks for security risks.

  4. 4

    You review and merge

    Every change moves to Review, where you read it, approve it or send it back with a note.

What lands in Review

  • A short spec with scope, non-goals and acceptance criteria
  • A ticket breakdown in small, ordered steps
  • Working scripts, dashboards or small internal apps built test-first
  • A bug fix with the root cause explained and a regression test
  • A code review with findings ranked as blocker, should-fix or nit
  • A security review with the impact and a concrete fix for each finding

What stays with you

  • Merging and deploying stay with you: agents propose changes, you approve them.
  • Shield never runs exploits against systems you do not own.
  • Design decisions are recorded with their options so you can overrule them.

Frequently Asked Questions About the Software Engineering Squad

Can AI agents build internal tools for my company?

Yes, for scripts, dashboards and small internal apps. In the Software Engineering Squad, Blueprint writes the spec, Forge builds it test-first and Check reviews it. The result lands in Review, and a person decides whether it is merged.

Which AI agents are in the Software Engineering Squad?

Five: Blueprint, the Tech Lead who writes specs and tickets; Forge, the Software Engineer who builds; Patch, the Maintainer who fixes bugs; Check, the Code Reviewer who flags risks before merge; and Shield, the Security Reviewer who hardens code and configs.

Can an AI agent review my code before I merge it?

Yes. Check reviews changes for correctness, readability and missing tests, and ranks each finding as a blocker, a should-fix or a nit. You still read the review and make the merge decision.

How does an AI agent fix a bug without guessing?

Patch reproduces the bug first, forms a hypothesis, tests it to find the root cause, then fixes it and adds a regression test. The fix and the explanation both land in Review.

Is it safe to let AI agents check my code for security problems?

Shield reviews input handling, auth, secrets, dependencies and data exposure, and explains each finding with its impact and a concrete fix. It never runs exploits against systems you do not own, and a person approves every change.

How is this different from an AI coding assistant in my editor?

An editor assistant helps one person with one prompt at a time. The Software Engineering Squad is five agents with separate roles sharing a task board, so specs, building, review and security checks stay visible as cards you can track.

See all 13 squads

Put the Software Engineering Squad to work

Start from this template as it ships, or have the xSquads team set it up for your team and tune the agents to how you work.