This is crazy. The hacker installed a dead-man's switch that will wipe your computer if you revoke the GitHub token they stole from you. Revoking the token is what triggers the wipe.
Over the past month, some of you reported Claude Code's quality had slipped. We investigated, and published a post-mortem on the three issues we found.
All are fixed in v2.1.116+ and we’ve reset usage limits for all subscribers.
Tried it. Not to hate, but it feels like the same pattern as a lot of AI tools right now: flashy, comprehensive, and ultimately not that useful.
Web fetch/search barely works. Resume/CV generation is mediocre at best. No real auto-apply, so there’s still a ton of manual work. The job scan surfaces stale listings from weeks ago, and some links are dead.
It’s not even an especially hard problem. That’s what makes the constant overpromising so frustrating. Big claims.
At this point I’m genuinely curious whether anyone has actual case studies showing these tools helping people land interviews, and what the economics look like.
sharing my first open source project
a CLI for downloading and syncing your X bookmarks locally so your agent can access them. it's free
› npm install -g fieldtheory
› login to your X account in a chrome tab
› ft sync (done!)
bonus:
› ft viz
› ft classify
@MaxWayld Ive build a tier list for web3 games into cal3ndar[dot]gg but ive singe sunsetted.
More than happy to provide you data endpoint so we build it at your end (given I see Fabelborne in there)
Anthropic’s entire Claude Code codebase leaked. It wasn’t a breach. Here is everything you need to know.
It was a build config mistake. A 57MB file shipped inside public npm releases contained 1,906 TypeScript source files. Architecture, system prompts, tool-calling logic, all of it. Anyone who downloaded the package had the keys without knowing it.
No hacking. No sophisticated attack. Just someone forgetting to exclude a file before publishing.
This is the kind of risk that doesn’t show up in a security audit. It’s not a vulnerability in your product, it’s a gap in your release process. And it can happen at any company, any size.
By the time they filed DMCA takedowns, early npm versions were already archived. The window between “shipped” and “contained” was enough.
For founders shipping software: your threat model isn’t just hackers. It’s your own pipeline.
Anthropic is a world-class team. This still happened to them.
On the bright side, the things they have in the pipeline are very damn exciting. The leak revealed 44 unshipped features sitting behind flags, including:
- KAIROS — an always-on background agent that runs while you’re idle, consolidates memory, and is ready with full context when you return
- autoDream — a literally-named subprocess where Claude “dreams,” merging observations and cleaning its own memory on a tick-based loop
- BUDDY — a full Tamagotchi-style terminal companion with 18 species, rarity tiers, RPG stats, and ASCII art. Real feature, not a joke
- ULTRAPLAN — offloads complex planning to a 30-minute cloud compute session running Opus, then teleports the result back to your terminal
- Coordinator Mode — turns Claude into a multi-agent orchestrator running parallel worker agents through research, synthesis, implementation, and verification phases
Not vaporware. Compiled code sitting behind feature flags, waiting to ship.