Development

A coding agent rewrites its own codebase every day

A vendor ships, as an ordinary product feature, the loop the newsletter is named for: the agent's own usage becomes the signal for the next day's version of the agent, on a fixed daily cadence rather than a release cycle.

development2026-01-27rsiself-modificationagent-harness

Claim

Factory AI released a coding agent that analyzes its own interactions and updates its codebase daily, while Anthropic introduced MCP Apps letting tools render interactive interfaces inside the chat.

Reporting date: (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Verification record

These entries record a check, not proof that the underlying claim is true. Unless described in the article below, the scope of the check is unknown.

  • Recorded by claude-fable-5-1/2026-09-17 on 2026-09-17T08:00:00Z

Factory AI’s Signals system analyzes the agent’s own interactions and turns what it finds into daily updates to the agent’s codebase, so the product that writes code is now partly written from its own usage (announcement). The same item records Anthropic’s MCP Apps, which lets tools render interactive interfaces inside the chat (MCP blog), filed by the newsletter under one heading: the recursive loop closing. It follows Claude Code’s Tasks two days earlier and the self-written Cowork app as scaffolding steps in the recursive self-improvement trajectory, and it precedes the emergent version four days later, when a Moltbook agent hardened its own loop after an SSH attack with no vendor involved.

Knowledge graphopen in browser →

Related to
Reported in
Actor
Evidence for
Supersedes
Referenced by

Cite or correct this record

Suggest a correction — opens a GitHub draft for you to review and submit.