Run an issue backlog with one agent per ticket and checks it did not write

A mobile developer moved from one-off Claude Code prompts to a harness that runs one agent per issue; Claude Code wrote most of its 643 commits and 96 merged PRs in under two weeks.

Done withClaude Code· Nov 16, 2026

What they did

Souleymane Sidibe, who mostly builds mobile apps, describes a year of working with Claude Code in stages. He added a CLAUDE.md per repo and custom commands, then chained planning skills (a questioning session, a spec, a split into issues, then implementation) so each step reads a file instead of a long chat. He wrote a stricter implement skill that demands real command output as evidence, tried a shell loop that fed the agent issue after issue, and then a manager skill that ran one background session per ticket in waves that avoid touching the same files, with locks on the shared emulators. That setup led him to build Loop, a Kotlin harness that pulls issues from a tracker, gives each one its own workspace and agent, and ends every run with a verdict from checks the agent did not write. Agent pull requests get a risk score from a cheaper model, and only low-risk ones merge without him.

How it went

He ran the implement skill 138 times before building the manager. Measured cost per issue had a median of 5 to 18 million tokens, one issue in ten went past 45 million and the worst passed 100 million, mostly runs nothing stopped. In one run every check was green but the agent had left uncommitted files, so Loop refuted it and sent the issue back. Since its first commit on November 5, Loop has 643 commits, 96 merged pull requests and about 2,270 tests, mostly written by Claude Code, and it now works on its own issues. Per-issue token budgets are not built yet.

Worth knowing

The self-run loop and the manager skill were fragile: the same agent wrote the code and the checks that called it done, and nothing capped spend. Weakened tests, red CI or diffs over 1,000 lines always go to a person. Loop's repo is private for now, so the numbers are the author's own.

Try it

With Claude Code· 5 parts to fill in

Work through the open issues labelled [ready] in [my tracker] for [repo], one issue at a time, each on its own branch. For each issue: read it, write a short plan, implement it, then run [test and build commands] and paste their real output into the pull request. Do not edit or weaken existing tests. Before opening the PR, confirm the work tree has no uncommitted files. Stop an issue and report back if it passes [token or time budget] or fails the checks three times. Open pull requests but do not merge anything; give me a list of PRs with their check output when the queue is empty.

Square brackets are the parts to fill in. The prompt is ours; the story is theirs.

Where they said it

Related