{
  "id": 11633822,
  "title": "Anthropic opens Claude Code to internal Mods",
  "url": "https://urgent.news/2026/10/03/anthropic-opens-claude-code-to-internal-mods",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-03T07:37:53.000Z",
  "source": {
    "name": "Arabian Post",
    "slug": "arabian-post",
    "url": "https://thearabianpost.com/anthropic-opens-claude-code-to-internal-mods/"
  },
  "original_language": "en",
  "account": "Anthropic has unveiled a new feature called Mods for Claude Code, empowering developers to customize the coding agent's behavior, user interface, and internal event handling using JavaScript or TypeScript functions running within the product. Introduced alongside Claude Code version 2.1.287, Mods build upon the existing plugin framework. These Mods can manipulate prompts before they reach the model, block or alter tool calls, approve or reject permission requests, redact data from tool output, and modify or augment parts of Claude Code's interface. Claude Code's Mods are packaged within plugins, working seamlessly in both the command-line interface and desktop application. Developers can either create their own Mods or request Claude Code to generate one, then install and reload it during an active session. This approach offers developers considerably more control compared to Claude Code's earlier hooks, which could trigger external actions at predefined points. However, Mods go a step further by executing functions in response to events emitted by Claude Code itself. Events available to Mods include tool calls, permission requests, and interface rendering, enabling developers to address a wide range of functions, from development safeguards to specialized displays. Anthropic has demonstrated various use cases for Mods, such as displaying continuous integration and deployment status alongside a conversation, requiring confirmation before commands that modify production configurations, and recording tool activity for audit purposes. The company has also integrated certain Claude Code functionalities into the new Mods architecture, such as the built-in /diff feature, which now functions as a Mod, allowing users to disable it or replace it with their preferred implementation. Anthropic plans to gradually migrate additional built-in capabilities to Mods, enabling users to maintain a smaller core and selectively add features as needed. Sample Mods showcase the versatility of this framework, ranging from lightweight interface enhancements like Token Weather, which displays context-window usage above the prompt, to more robust control features like Blast Radius, which can prevent potentially risky shell commands and show their potential impact before the user decides whether to proceed. Replay Theater is another example, recording edits made during a turn and allowing developers to review them sequentially. While this newfound flexibility offers significant advantages, it also introduces a critical security consideration. Anthropic emphasizes that Mods are not sandboxed and run with the same access to a computer as Claude Code, meaning they have unrestricted access to a developer's machine. Users are advised to install Mods only from trusted sources, treating them as other code executed on their machines. To help maintain centrally managed restrictions, Anthropic has introduced security measures for Team and Enterprise deployments, as well as machines using managed settings. A built-in security Mod called sec-default is loaded first in these environments, preventing user-installed Mods from overriding protected organizational settings. Administrators can also control which plugin marketplaces users may access and can even load their own Mods before user extensions. Mods require Claude Code 2.1.287 or later and are enabled by default in supported versions. Developers are cautioned that the Mod programming interface can change between releases, as Anthropic provides TypeScript declarations specific to each version. The introduction of Anthropic's Mods for Claude Code represents a significant step forward in customizing and securing the coding agent, providing developers with unprecedented control and flexibility while maintaining robust security controls for enterprise environments.",
  "summary": "Anthropic has introduced Mods for Claude Code, giving developers a way to alter the coding agent’s behaviour, interface and internal event handling through JavaScript or TypeScript functions running inside the product. Claude The system, released on October 1 with Claude Code version 2.1.287, extends the tool’s existing plugin framework. Mods can rewrite prompts before they reach the model, block…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 2,
    "also_reported_by": [
      {
        "outlet": "Dev.to",
        "title": "Turning Claude Code into a Writing Harness with Claude Mods and an Output Style",
        "url": "https://urgent.news/2026/10/03/turning-claude-code-into-a-writing-harness-with-claude-mods-and-an",
        "published": "2026-10-03T08:42:21.000Z"
      }
    ]
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}