ALTK-Evolve: On-the-Job Learning for AI Agents

The ALTK-Evolve memory system turns AI agent trajectories into reusable rules. On the AppWorld benchmark, success rate improved by 8.9 percentage points on average and 14.2 points on hard tasks.

• Average SGC gain on AppWorld: 8.9 pp (from 50.0% to 58.9%)
• Hard task success rose 14.2 pp to 33.3%
• System extracts rules from trajectories and injects only relevant ones
• No-code plugin for Claude Code and low-code ReAct agent integration