Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-08-19 · high
CLAUDE_CODE_ENABLE_TODO_TOOLS=1 fixes the August 15 removal from Opus 4.8, Sonnet 5, Fable 5, and Mythos 5+, but nothing flags skills that need it.
Why now: code.claude.com's changelog through 2.1.235 confirms the fix, closing out a week of environment-variable changes covered here since August 15.
CLAUDE_CODE_ENABLE_TODO_TOOLS=1 restores the todo tool pulled from Opus 4.8, Sonnet 5, Fable 5, and Mythos 5+ on August 15, per Claude Code's changelog through 2.1.235.
That removal broke anything relying on the model to track multi-step work through a dedicated tool. The fix is one environment variable, but nothing signals when you need it.
Other changes landed in the same window. Bash got memory cgroups on August 15. WebFetch picked up a cache TTL and a forward_user_identity flag on August 17. Per-project config directories and quota-reset auto-resume landed on August 18. Those are additions, not removals, so the changelog doesn't say they carry the same failure risk.
The todo removal is the one that bites without warning. If a skill or prompt still references a todo or task tool, it won't throw an error. The model just stops tracking steps the way the skill expects, and it reads like the model lost the thread on its own.
Grep your skill files for todo or task tool references before you blame the model for dropping a long task. If you find one, set CLAUDE_CODE_ENABLE_TODO_TOOLS=1 and rerun it.
Each link below shares sources, entities, or timing with this story.
Claude Code 2.1.233 added an environment variable configuring the WebFetch session URL cache TTL, with the default unchanged at 15 minutes (Claude Code Changelog). That default is a real problem for any agent polling a status page, a CI endpoint or a changelog inside a long se...
Buried in the 2.1.233 notes. The harness is removing explicit planning scaffolding for the newer tier while keeping it for older models, betting these models track multi-step work internally and that a forced todo list burns tokens without improving completion. Claude Code Rel...
Fable 5 and Mythos 5 went dark this week. Not a soft sunset with a six-month migration window. A US export directive, and within hours the models were unavailable to any foreign national anywhere on earth. Enterprise teams outside the US woke up to API calls failing against a...
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...
Terminal-Bench 2.1 results (entries dated June 17) put Codex CLI on GPT-5.5 first at 83.4%, Claude Code on Fable 5 second at 83.1%, and Claude Code on Opus 4.8 at 78.9%. The asterisk matters more than the ranking: Fable 5 and Mythos 5 have been export-suspended since June 12,...
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...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.