Anthropic introduces Claude Code mods

Anthropic has introduced Claude Code mods — TypeScript functions that intercept agent events, rewrite prompts, block tool calls, manage permissions, and strip secrets from output.

Compiled by Sergey KostenchukPublished 2026-10-03Updated 2026-10-03

2026-10-03 Coding Anthropic

🛠 Anthropic introduces Claude Code mods

The Machinelearning digest for October 3 covered five events. The main one: TypeScript mods intercept Claude Code events, rewrite prompts before they are sent to the model, block tool calls, and strip secrets from output. The built-in /diff has been rewritten as a mod.

🌍 The agent runtime has become a platform: mods don't just observe like old hooks, they rewrite behavior. There is no sandbox — access to the machine at the Claude Code level; on Team and Enterprise, risks are addressed by the built-in security mod sec-default.

👤 Mods are installed via /plugin or the Claude catalog; the assistant writes a mod for itself — it is picked up without restarting the session. Only install trusted code. Also in this issue: FLUX 3 Image (from $0.0205 on OpenRouter) and MAI-Transcribe-2-Streaming — transcription in 60 languages for $0.54/hour.

Source 1: https://claude.com/blog/claude-code-mods Source 2: https://microsoft.ai/news/our-first-streaming-transcription-model/

Sources