List your company

Task-oriented Agent Skill Evolution on SkillFlow

100Cross-Format Data Reconciliation ScoreSelf-Evolve
7480.7587.594.25Jun 2, 2026

Evaluation results

MethodLinks
Self-EvolveHeterogeneity Setting=Backbone, Backbone=Qwen3.6-Plus, Harness=Claude Code2026.0610062.587.55062.522.262.5500502566.737.566.7036.75
FederatedSkillHeterogeneity Setting=Backbone, Backbone=Qwen3.6-Plus, Harness=Claude Code2026.06100751005087.533.387.55011.15037.577.837.577.8043.37
FederatedSkillHeterogeneity Setting=Backbone, Backbone=Kimi K2.5, Harness=Claude Code2026.06100507562.587.522.21005011.17537.544.42588.9040.96
Self-EvolveHeterogeneity Setting=Backbone & Harness, Backbone=Qwen3.6-Plus, Harness=Qwen Coder2026.0610062.5100507566.710037.522.2755077.82577.8045.78
FederatedSkillHeterogeneity Setting=Backbone & Harness, Backbone=Qwen3.6-Plus, Harness=Qwen Coder2026.06100751005010055.610062.522.262.537.566.75088.912.548.8
FederatedSkillHeterogeneity Setting=Backbone & Harness, Backbone=GLM-5, Harness=Claude Code2026.06100501005087.511.11005011.17537.566.737.5100043.37
Self-EvolveHeterogeneity Setting=Backbone, Backbone=GLM-5, Harness=Claude Code2026.0687.562.587.55087.5010037.5062.52533.325100037.35
FederatedSkillHeterogeneity Setting=Backbone, Backbone=GLM-5, Harness=Claude Code2026.0687.562.587.5507522.210062.507537.566.750100043.37
Self-EvolveHeterogeneity Setting=Backbone, Backbone=Kimi K2.5, Harness=Claude Code2026.0687.5757537.562.511.187.562.522.262.52511.112.5100036.14
Self-EvolveHeterogeneity Setting=Backbone & Harness, Backbone=GLM-5, Harness=Claude Code2026.0687.562.587.55087.5010037.5062.52533.325100037.35
Self-EvolveHeterogeneity Setting=Backbone & Harness, Backbone=Kimi K2.5, Harness=Kimi CLI2026.0687.537.587.537.587.533.310062.5062.537.544.412.588.9038.55
FederatedSkillHeterogeneity Setting=Backbone & Harness, Backbone=Kimi K2.5, Harness=Kimi CLI2026.067562.510037.587.566.71005011.162.52588.912.5100043.98