AgentLayer▸docs
Engineering

Engineering

How Kevin works on code. A playbook per kind of task, 23 named principles, and review, verification, and release loops that never touch your remote on their own.

Kevin works on code the way a careful senior engineer does. It reads the repo's manual first, picks a playbook for the kind of task in front of it, proves the result on the running artifact, and hands you a diff with the comments already stripped. Run a skill as /agent-kevin:<name> in Claude Code or $<name> in Codex; these docs use the bare name.

The loop, in order

  1. Plan. plan-spec turns an intent into a spec with a design section and a check on every task. See Specs and plans.
  2. Isolate. setup-worktree gives each stream of work its own bootstrapped checkout. See Worktrees.
  3. Build by the playbook. engineer routes the task (bug fix, feature, refactor, performance, prototype, and more) and applies the principles as it goes. See The engineer skill.
  4. Prove it. The result is checked on the real surface, not inferred from a green build. See Verification.
  5. Review. A teammate's PR gets verified findings with paste-ready comments, the review on yours gets answered, and engineer preps you to present your own. Your own work, code or not, gets a second model with adversarial-review. See Pull requests and Second-model review.
  6. Release. release cuts a version for maintainers; upgrade brings every home current. See Releases.

Kevin never pushes, tags, releases, or opens a PR on its own. Commits happen only in the scope you approved, and history is forward-only.

On this page