Fetching from the wire…
Public story · 2026-07-22 · high
The July 19 update made /verify and /code-review opt-in, so review no longer fires automatically and nothing warns you it's off.
Why now: The change landed July 19, two days before a Stack Overflow post argued agentic engineering depends on exactly the verification layer this release turned off by default.
Claude Code stopped auto-running /verify and /code-review in the July 19 release, v2.1.215, per the changelog.
If your workflow assumed a review pass fired automatically at the end of a task, it doesn't anymore, and nothing warns you. Code that used to get checked ships unreviewed by default.
Fix it with a hook on task completion, or an explicit /code-review call in your task template. Or add a CLAUDE.md line telling the agent to run it before declaring done.
I'd take the hook. A memory-file instruction is subject to the same silent instruction-dropping an 80-instruction study already documented.
This lands inside a run of harness bugs that share one shape.
Workflow saves followed a symlink planted at .claude and wrote outside the project tree. claude daemon stop --any used stale lockfiles to pick targets and could kill unrelated processes.
Truncated MCP outputs kept the full untruncated payload resident in memory, leaking steadily behind a tidy truncated display. Auto-compact never fired at all for Opus 4.8 on Bedrock, per releasebot.
None of that is a model problem. Every one is the harness disagreeing with reality.
What the sandbox thinks a path is versus what the filesystem does. What the display shows versus what memory actually holds.
Four shell-parser permission fixes shipped in the same release, including PowerShell validation of commands with invisible Unicode. That's the profile of an injection payload that reads safe to a reviewer but executes differently in the shell.
Two fixes are worth the upgrade by themselves. v2.1.217 warns when transcript writes fail or session saving got disabled by an inherited env var. That's what bites headless -p runs launched from cron or launchd.
v2.1.216 fixed message normalization scaling quadratically with turn count, the cause of multi-second mid-session stalls.
Stack Overflow named this shift two days later: the move from vibe coding to agentic engineering, a matter of verification, review, and rollback scaffolding. That's the layer that shipped four bugs in the same release.
Each link below shares sources, entities, or timing with this story.
Two researchers in my set surfaced this independently, which is usually a sign it matters. Claude Code 2.1.166, first seen June 6, introduces a fallback-models setting: configure up to three models tried in order when the primary is overloaded or unavailable. It also adds glob...
This one rearranged my week. An essay published August 4 walks through Databricks' independent benchmark of coding harnesses against its own multi-million-line codebase. Pi, a harness with four built-in tools and a system prompt under 1,000 tokens, paired with Opus 4.8 at xhig...
xAI launched Grok Build on May 14. With that, every major AI lab now ships a coding agent that lives in your terminal. The competition isn't "can we build one" anymore. That question is settled. The lineup: Anthropic has Claude Code. OpenAI has Codex CLI. Google has Gemini CLI...
Go rotate a key. I'll wait. Claude Code 2.1.246, released August 25, lists this in its changelog: a fix for "telemetry and metrics requests to Anthropic carrying the API key configured for a third-party gateway (ANTHROPIC_BASE_URL); a credential is now only sent to its own hos...
Two releases landed today and they're both significant. The headline security fix: PreToolUse hooks returning allow were bypassing deny permission rules — including enterprise managed settings. If you're running Claude Code in a managed enterprise environment with security hoo...
It now dedupes local CLAUDE.md files against checked-in ones, proposes trimming content it can derive from the codebase, ranks unused skills, MCP servers, and plugins by context cost, and flags slow hooks. Background subagents run by default and, when launched from claude agen...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.