FACET: Building Executable Terminal Tasks from Agent Skills

_akhaliq · x · 2026-08-22

FACET is a method for building executable terminal tasks from agent skills while preserving intent and state. It reconstructs scenarios and builds the environment first, then aligns instructions, solutions, and verifiers to that state. The approach produces 6,078 validated tasks with dense checks. Fine-tuning with 1.2K trajectories improves performance on Terminal-Bench 2.1.

Related event: FACET: Synthesizing Executable Terminal Tasks While Preserving Intent(2 posts)→

Original post →

More from coding & agent

coding & agent channel →