Claude Blog · Oct 17, 2026 · 2:28 AM

How the Claude Code team reviews code: risk tiers, a 3-round cap, and zero-human merges for low risk

Anthropic's Oct 15 write-up says merged Claude Code PRs more than doubled from June to September while the revert share fell over 40%, with most reading moved to Claude reviewers, blocking security review, and classifier-set risk tiers where low-risk changes from trusted authors need no human approval. After an agent-vs-agent PR looped through 50+ overnight rounds, the reviewer now stops after three rounds and re-reviews only the new diff.

Read at claude.dev

Claude Code

An AI coding agent for terminal, IDE, web, and Slack

More news

View all

Postgres 19 slips to Nov 19 after LLM-assisted review surfaces more bugs than committers can fix

PostgreSQL 19 GA moves to November 19 (RC2 on Nov 12) after 8 to 10 significant features, including SQL/PGQ property graph queries, were reverted to v20 during beta. Committer Peter Eisentraut says LLM assisted code review found long tails of edge case defects faster than volunteers could triage them, and stresses the reverts weren't caused by vibe coded features…

The Register 55 minutes ago

Reflection releases Beam weights under Apache 2.0, plus its Mirror CLI coding harness

As of Oct 16, Beam (501B MoE, 23B active, 1M context) is on Hugging Face as reflection/Beam 1.0 501B A23B under Apache 2.0, with day 0 Transformers, SGLang, and vLLM support and a free to start API at platform.reflection.ai. Reflection also shipped Mirror CLI in beta, the terminal harness Beam was trained with, which runs parallel agents in separate worktrees and supports MCP, skills, and tool call approval…

Reflection Blog 3 hours ago

Pydantic AI 2.57 adds ApprovalReviewer: a model auto-approves low-risk tool calls

Pydantic AI v2.57.0 (Oct 15) adds Harness ApprovalReviewer , which asks a second model one risk question about each tool call marked requires approval and runs it on a "no", while "yes", "unsure", or a model failure still goes to a person. It never denies or runs calls itself, and pairs with ToolCallJudge when calls should be blocked…

GitHub (pydantic/pydantic-ai) 4 hours ago

OpenAI sets Jan 14, 2027 API shutdown for GPT-4o, 4o mini, GPT-4.1, and 4.1 mini

OpenAI's deprecations page says the gpt 4o 2024 08 06 , gpt 4o 2024 11 20 , gpt 4o mini 2024 07 18 , gpt 4.1 2025 04 14 , and gpt 4.1 mini 2025 04 14 snapshots leave the API on January 14, 2027, with gpt 6 luna as the recommended replacement for all of them. Fine tunes built on these models shut down on the same date, so agents still pinned to them have about three months to migrate and re test tool calls…

OpenAI API Docs 5 hours ago

Composio adds approval checkpoints before agent session tools run

As of Oct 16, Composio sessions can require a person's approval before selected toolkits, tools, or tagged tools execute: set requireApproval ( require approval in Python), catch the input required error, show its requests, then resend the call with the answers. It covers meta tool calls too, and over MCP it uses elicitation, with a deny or allow fallback for clients that can't ask…

Composio Changelog 6 hours ago

Spotted something we missed? Start a thread.