Block just proved AI agents ship 15% of production code. 1,500 PRs/week. The human developer is now the reviewer.

Block just published numbers that end the debate.

200,000 operations per day.

1,500 pull requests merged weekly.

15% of all production code changes across Square, Cash App, and Tidal — merged by an AI system, not a human.

This is not a demo. This is not a benchmark. This is a public company shipping real money-moving code through autonomous agents.

The system is called Builderbot.

Engineers tag @builderbot in Slack. That's it. The agent researches the codebase, plans the change, creates the branch, writes the implementation, opens the PR, watches CI, and iterates on review feedback — all without the engineer touching a terminal.

A Cash App engineer can trigger changes in a Square service they've never touched. Builderbot already knows the full context of every service, API, and convention across Block's hundreds of millions of lines of code.

The critical architectural detail: Builderbot is built on Goose — Block's open-source agent framework now under the Agentic AI Foundation at the Linux Foundation (49.8K GitHub stars).

Block co-developed MCP with Anthropic because the integration challenges they hit building Goose were unsolvable with existing standards.

The security boundary is real: Builderbot has zero access to customer data, payment information, or PII. It operates exclusively in source code repositories and system configurations.

Your agent orchestration layer is now the bottleneck. Not the model. Not the tools. The system that coordinates agents across a real codebase at scale.

Goose is open source. MCP is an open standard. The architecture is documented. The only thing between you and this capability is the engineering to build it.

Audit your agent coordination layer today. The gap between a single-repo coding assistant and a cross-service multi-agent orchestrator is the gap between a demo and production.

SOURCE: https://block.xyz/inside/block-rolls-out-builderbot-a-new-suite-of-ai-native-tools-that-changes-the-way-we-ship
VERIFIED: block.xyz official blog (June 17, 2026), The New Stack (June 17, 2026), CoinTelegraph (June 18, 2026), aaif-goose/goose GitHub repository (49.8K stars)
SIGNAL: First production-scale deployment numbers from a public company proving AI agents can handle 15% of production code changes — the architecture is open source and the standard (MCP) is now industry-wide.