Anthropic released Claude Code version 2.1.294, addressing specific issues with prompt and agent hooks. The update fixes a bug where instructions intended to block certain commands were not being recognized as valid blocking criteria. Additionally, the release improves how stop conditions are evaluated for subagents, reducing instances of premature termination. These changes enhance the reliability of automated coding tasks that depend on precise instruction following.
Read originalEarlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
Duncan Rogoff · June 19, 2026 · Related
Duncan Rogoff · July 28, 2026 · Related
Lev Selector · July 31, 2026 · Related
Claude Code Releases · September 16, 2026 · Same story
The Verge AI · September 17, 2026 · Related
Claude Code Releases · October 6, 2026 · Same story
This release tightens the leash on Claude Code's autonomous capabilities while fixing critical sandbox escapes. The new effort parameter for Agent tools lets developers explicitly control sub-agent depth, a necessary guardrail as these systems grow more complex. Security fixes are prominent, addressing how plugins handle network paths and how file permissions persist during session resumption. It’s a stability patch that ensures the tool remains usable in enterprise environments without compromising on the new agent features.
Anthropic quietly shipped a significant model update alongside routine maintenance. Claude Haiku 5.5 is now the default on the API, offering a 1M context window at $0.10 per million tokens, which lowers the cost floor for high-volume coding tasks. The release also patches critical stability issues in the local agent runtime, specifically fixing memory leaks in HTTP MCP connections and resolving session state corruption during context compaction. These fixes matter because they stabilize the autonomous coding workflow that developers rely on daily. With Haiku 5.5 now standard, teams can deploy cheaper, faster iterations without manual configuration.
This release prioritizes stability over new features, addressing critical reliability issues in Claude Code's backgrounding and MCP integration. The most significant change is the addition of an 'onFailure: block' for hooks, which prevents silent failures from bypassing safety checks—a crucial update for developers relying on automated workflows. Gateway connectivity also sees major improvements with upstream timeout controls and better error handling for Bedrock and Vertex integrations. While there are no headline-grabbing capabilities, these fixes make the tool significantly more robust for heavy daily use.
This release adds a critical observability layer to vLLM's KV cache system by exposing prompt token counts broken down by cache tier. For operators running large-scale inference, seeing exactly how much data is served from the GPU versus CPU or disk is essential for tuning memory allocation and cost efficiency. It transforms the cache from a black box into a measurable metric, allowing teams to validate whether their prefill strategies are actually hitting the intended storage layers. This granularity helps prevent over-provisioning and identifies bottlenecks in prompt processing pipelines.
This release quietly fixes a critical accuracy gap for ModernBERT encoders by implementing exact GELU activation, ensuring semantic embeddings match the original PyTorch models rather than approximations. It also brings native support for CUDA 13.4 across Linux and Windows, closing the driver compatibility lag that has plagued NVIDIA users on newer hardware stacks. While KleidiAI builds are temporarily disabled on Apple Silicon, the broader expansion to ROCm 10.0 and Snapdragon NPU keeps llama.cpp as the most versatile local inference runtime available today.
This release targets a specific but painful stability issue for Android users running llama.cpp on Qualcomm Adreno A6X GPUs. The kernel compiler was crashing due to argument limits in the iot device backend, effectively breaking local inference on those chips. By skipping the problematic kernel and adding explicit detection for the Adreno 623, the team restores functionality where it previously failed hard. It’s a narrow fix, but essential for anyone trying to run models on mid-range Android hardware without hitting compiler errors.