Click any tag below to further narrow down your results
Links
Ray Amjad argues that "Code Mode," a new way of running tools inside Claude Code, is the largest change Anthropic has made to Claude Code's harness so far. The video walks through how it differs from the standard tool-calling loop and why that matters for agent behavior.
- The change concerns the harness, the scaffolding around the model that decides how tools are called and results are returned, rather than the model itself.
- Instead of the model making one tool call at a time and reading each result back, Code Mode has it write code that calls the tools, so several steps run in a single pass.
- The video's case is that this cuts down on context bloat and round trips, which is why he calls it the biggest harness change yet.
The article argues that agent harnesses are collapsing large tool catalogs into a single exec(code) tool, where the model writes a script that chains actions in a sandbox instead of calling tools one turn at a time. It covers why this helps and what it changes about how tools need to be designed.
- Standard tool calling is one action per turn, so chained workflows burn round-trips and context; Code Mode lets the model script dependent steps, loops, and parallel calls in one pass.
- Only the final summary returns to the context window; intermediate payloads stay in the sandbox, which the author ties to context savings and cost reductions (Factory reported a 15% drop in model spend and a +3.5 point Terminal-Bench gain after switching Droid to Code Mode).
- Typed interfaces, a credential and policy bridge, and persistent state outside the prompt make a JS or Python sandbox more practical than bash, though existing CLIs and skills may need rework to fit.