Anthropic proposes Function Hooks to let developers deeply customize Claude Code
EricBuess · x · 2026-09-04
- Anthropic's Claude Code team (@ClaudeDevs) shared an early proposal for Function Hooks, a not-yet-shipped extensibility mechanism, soliciting feedback via GitHub issue #91870.
- Core idea: developers could hook into Claude Code with TypeScript functions à la Express/Koa middleware and change almost anything — including how components render (CC uses React internally) — composed via a registration-order next continuation model, with safety from side-effect tracking on a parameterized $ object.
- Workplace admins could also use hooks to control team behavior at a fine programmatic level, including restricting what changes others can make.
- The team linked a technical architecture doc and demo videos; the feature hasn't shipped and feedback is open on GitHub.
More from coding & agent
- Clef on Ollama: full API examples for Cloudflare's 27B decision model — ollama · 2026-10-21
- Opus 5.5 distills a 30-min pi-durable lecture into a 7-min narrated explainer video — solyarisoftware · 2026-10-21
- OpenSearch MCP Server lets agents query and analyze Wazuh security logs — modelcontextprotocol · 2026-10-21
- chaos-index MCP server wraps StupidAPIs with no API key required — modelcontextprotocol · 2026-10-21
- UC Berkeley open-sources SkyDiscover, an AI framework that evolves better algorithms and builds full systems — CShorten30 · 2026-10-21
- ruff author on ts-rust's trick: vendor and trim forked deps for a language subset — charliermarsh · 2026-10-21