highAIOctober 1, 2026
Anthropic Adds Mods to Claude Code, Turning the Agent Into a Programmable Runtime
Claude Code mods, released October 1, 2026, are TypeScript or JavaScript functions that can rewrite prompts, gate tool calls, redact secrets and replace parts of the interface.
Source: AI Weekly
Pulse Take
Mods turn a coding agent into a platform, much as extensions did for browsers and editors. The upside is enforceable team policy — secret redaction and permission rules in code instead of in prompts. The risk is the same as with any extension: mods run with Claude Code's full access, so treat them like dependencies.
Event
Anthropic released mods for Claude Code on October 1, 2026. Mods are JavaScript or TypeScript functions that hook into the agent's execution pipeline. They are distributed inside plugins and installed with the
/plugin command in the CLI or desktop app, and developers can write one themselves or ask Claude to build it.Impact
A mod can rewrite a prompt before it reaches the model, block or retry a tool call, approve or deny a permission request, redact secrets from tool output, and edit or replace interface elements. Anthropic says it built features such as
/diff and AGENTS.md support as mods. Because mods run with the same access to your machine as Claude Code, Anthropic advises installing them only from trusted sources.What to do
Teams can use mods to enforce policy in code — for example redacting credentials or blocking destructive commands. Vet third-party mods as you would any dependency. Compare coding agents in Claude Code vs OpenAI Codex and review our MCP security playbook.
Related Playbook
View Playbook →