An agent working our backlog.
We don't sell vapor: we run what we build. This is our own agent, working our own Jira, with real numbers.
163
autonomous cycles in 3 weeks
The numbers
163
autonomous cycles in 3 weeks
32
PRs reviewed, 4 review follow-ups
13
backlog tickets closed
190s → 79s
median cycle time after optimization
The loop, live
- 01
Jira backlog
prioritized tickets
- 02
Claude Code agent
autonomous cycles
- 03
PR + tests
never straight to main
- 04
Automated review
approve if only nits
- 05
Merge + evidence
logs, diffs, decisions
The rules that made it work
Human-in-the-loop by design: the agent escalates whatever needs judgment; it doesn't improvise.
"Approve if only nits" — and "nit" is defined in writing. Without a definition, the rule is a roulette.
Everything goes through PRs with tests, never straight to the default branch. Autonomy lives inside guardrails.
Every cycle leaves evidence: logs, diffs, decisions. If it can't be audited, it doesn't ship.
The stack
- Claude Code
- Jira
- GitHub
Claude Code + Jira + GitHub, orchestrated with our own evals and scheduling framework.
This is exactly what we set up for clients.
Agents with evals, guardrails and evidence. If you want one working your backlog: