Blog

/dr-verify: Three Different Ways to Challenge a Result

A deterministic floor, an isolated peer review, and native runtime dispatch produce different evidence instead of repeating the same check three times.

Read more →

Datarim 2.0: One Workflow, More Than One Runtime

Version 2.0 took Datarim beyond Claude without splitting its repository, workflow, or auditable instructions.

Read more →

A Self-Learning Loop That Cannot Approve Itself

Datarim turns repeated lessons into proposals and backlog work, while a human remains the authority for changing framework rules.

Read more →

Symlink-First Installation: The Runtime Is the Checkout

Datarim links runtime instructions to the git repository by default, removing a hidden copy that used to drift between edits and updates.

Read more →

Datarim 1.15: Research Phase — Know Before You Build

Version 1.15.0 adds a structured research phase to the pipeline. A new researcher agent investigates library versions, best practices, and security advisories before planning begins — and fills knowledge gaps during implementation.

Read more →

Datarim 1.11: General Cleanup Before Long Term Memory

Version 1.11.0 brings multi-task disambiguation, automated runtime-to-repo sync, optimized descriptions, and a full site refresh — setting the stage for LTM integration.

Read more →

Smart Model Routing in Datarim 1.6: Right Model for Every Agent

Datarim 1.6 introduces model assignment — each of the 16 agents and 14 task-skills now runs on the right model for its role. Opus for critical reasoning, Sonnet for standard work, Haiku for structured tasks.

Read more →

What is Datarim? An AI Workflow Framework for Claude Code

Introducing Datarim — a structured iterative workflow framework that turns any project into a series of managed tasks with AI assistance at every stage.

Read more →