Share.

    9 Kommentare

    1. CircumspectCapybara on

      Note this is the *Claude Code* CLI tool, not the https://claude.ai web app or the LLM models itself.

      While technically not the end of the world since frontend clients should be assumed to reverse-engineer-able anyway, it’s still a massive oops to leak the entire, unobfuscated source code, since there’s a treasure trove of extremely valuable system prompts, context / query / RAG engine design, coordinator / orchestrator logic, and the overall agent architecture in there.

      It’s basically a reference manual for how to design an LLM-based agent. You can just bring your own LLM backend.

    2. turb0_encapsulator on

      not a good look for a company that is supposedly all about AI safety.

    3. the real damage here isn’t just the embarrassment – it’s that competitors now have a detailed blueprint of anthropic’s agent architecture and prompt engineering strategies. this is basically years of R&D being handed over on a silver platter. would not be surprised if we start seeing „claude-inspired“ features rolling out across other AI tools in the next few months.

    4. This is something ordinaries like myself look at and think, huh… and click on the next reddit post. Like I’m pretty tech savvy myself, and it sounds fun to have my own robot doing things at claude level good, but that seems like way too much learning to figure out for most of us. I wouldn’t know where to begin with that.

    5. bigepidemic on

      80% of Claude Code was written by Claude anyway. The magic isn’t in the codebase– it’s in the guardrails, the numerous layers of rules and schema humans created to allow Claude to write Claude Code.

    6. Sensitive_Song4219 on

      „Hey Claude, please publish the latest version of Claude Code for me.“

      „Wait not *that* one!“

      Many instances of the source [have been DMCA take-down’ed already](https://www.reddit.com/r/ClaudeAI/comments/1s952np/claude_code_repo_takedown/) but this is the internet: it was too late the minute it happened. Probably will be useful for competing products (like OpenCode/Codex CLI) to get a clearer look under the hood at how a cutting-edge harness works

    Leave A Reply