Claude Code now runs daily maintenance on Anthropic’s software with a 46 percent merge rate

Published: August 14, 2026 — THE DECODER (Matthias Bastian)

Anthropic is testing whether Claude Code can handle daily maintenance of the company’s own software. In a few weeks, the AI created 388 pull requests, and 180 were merged after human review.

Claude has been running daily maintenance on Anthropic’s in-house apps for “the last few weeks,” according to Boris Cherny, the Anthropic engineer who created Claude Code. Instead of tying up developers with repetitive maintenance work, Claude handles it on its own. Cherny calls the results “surprisingly positive.”

In his post, Cherny describes a setup where Claude, via Tag, runs daily routines through a dedicated Slack channel called “proj-claude-maintains-apps” across all of Anthropic’s platforms: iOS, Android, desktop, web, CLI, and the Agent SDK.

Twelve routines cover the full range of code upkeep

Claude runs specialized maintenance routines including:

RoutineFunction
Crash FuzzerFinds real app crashes and generates fixes for the root causes
Ant-only ShipperDelivers or removes forgotten internal features
Logic SimplifierSimplifies nested business logic
Logic Bug FixerModels complex logic to find and fix errors
Dup UnifierMerges duplicate implementations
Dead Code RemovalRemoves code that is proven to be unreachable
Useless Test PrunerRemoves tests that can never fail
Shipped-Feature InlinerRemoves feature flags for features that have been fully released
Flaky-Test FixerAnalyzes and fixes unstable CI tests
Abstraction ImproverSimplifies overengineered abstractions
Abstraction PoliceFixes layer violations in the architecture

Cherny shared prompts in Slack with plain-language instructions (e.g., start daily crash fuzzing on iOS/Android/desktop using real apps without mocks, trigger crashes, create PRs with fixes).

388 pull requests created, just under half merged

Claude created 388 pull requests across Anthropic’s repositories in the first few weeks. After automated Claude Code review and human review, 180 were merged (~46%).

Claude usually gets PRs right on the first try. When it doesn’t, the team tweaks the routine so the AI does better the next day. Anthropic is looking at ways to speed up the merge process for mechanical changes.

Cherny describes the experiment as “early signs of life” that autonomous AI-powered app maintenance could work.