Claude-Code v2.1.30 Released with PDF and OAuth Enhancements
Claude-Code's latest release, v2.1.30, brings notable enhancements and fixes that developers should be aware of. This version introduces a new parameter for the Read tool, allowing specific page ranges to be extracted from PDFs, like pages: "1-5". Additionally, handling of large PDFs (over 10 pages) now returns a lightweight reference instead of embedding the content directly, improving context efficiency when mentioned.
For developers integrating with MCP servers that lack Dynamic Client Registration, the update offers pre-configured OAuth client credentials. Utilize the --client-id and --client-secret options when adding with claude mcp add.
The release also includes updates for debugging with the new /debug command, support for additional git log and git show flags in read-only mode, such as --topo-order and --cherry-pick, and improved visualization in VSCode with multiline input support for question dialogs (Shift+Enter for new lines).
Key bug fixes in this release address phantom "(no content)" text in API history, prompt cache invalidation issues, and 400 errors after using /login. Additionally, there's a 68% reduction in memory usage for sessions thanks to optimized stat-based loading.
This update is valuable for developers looking to efficiently handle large document processing and improve integration with third-party MCP services. Users experiencing issues with session handling or permissions will also find these improvements beneficial.
📖 Read the full source: GitHub Claude-Code
👀 See Also
Developer Replaces $25/hr Virtual Assistant with AI Agents, Confronts Ethical Implications
A developer replaced a $25/hour virtual assistant with AI agents that handle follow-ups, scheduling, lead tracking, and CRM updates. The AI setup costs about $1,000/month and performs tasks faster and more consistently than the human assistant.
Don’t Use AI to Write Things You Present as Your Own Work
James Bach argues against using AI to draft any content you claim as your own. He warns that admitting AI help devalues your reputation and treats all such work as slop.
Apple's libibverbs Hides GPUDirect RDMA Symbols; Zero-Copy Metal Buffer RDMA Works on macOS
A developer discovered that Apple's RDMA subsystem accepts Metal GPU buffers for zero-copy network transfers and found hidden ibv_reg_dmabuf_mr symbols suggesting GPUDirect RDMA is possible on macOS without kernel modification.
Kimi K2.6 beats Claude, GPT-5.5 and Gemini in coding challenge with aggressive sliding strategy
In the AI Coding Contest's Day 12 Word Gem Puzzle, Moonshot AI's open-weights Kimi K2.6 scored 22 match points (7-1-0), outperforming GPT-5.5 (16), Claude Opus 4.7 (12), and Gemini Pro 3.1 (9). MiMo V2-Pro took second. Kimi won by sliding aggressively.