🧰 ALTK-Evolve

The items that name it 2

• ALTK-Evolve: How learned memory tops the τ²-bench leaderboard on customer service tasks without a frontier model
  writeup2026-10-20

  Library that analyzes an agent's training-task attempts and uses an LLM to distill successes and failures into short behavioral guidelines, which are appended to the agent's system prompt. It needs no weight updates or human annotation.

• Your Agent Aced the Task. Will It Do It Again?
  article2026-09-15

  System that turns an agent's own past trajectories into reusable guidelines, distilled automatically and injected back into the agent's context at inference time; its open-source repo now includes the Consistency Analyzer and consistency-guideline generation.