Repo-To-Skill (arXiv 2609.02749, Thu 3 Sep digest) distills the operational knowledge stored in repositories and papers into reusable, verified skills. The DisCo agent produces them in two forms: task-agnostic — yielding the AREX-Skill library of 5,000+ verified skills from 1,000 widely-used ML repositories, organized into 20 domains and 178 capability families — and task-oriented, generated for the task at hand. With a fixed GPT-5.5 backbone, test harness, and execution budget, skill-equipped agents score +134.3% on MLE-bench, +34.4% on PaperBench, +9.2% on FrontierCS, and +14.0% on PassNet.
Repo-To-Skill distills GitHub repositories into verified agent skills
2026-09-04Frontier watchTRIAL
Repo-To-Skill (arXiv 2609.02749, Thu 3 Sep digest) distills the operational knowledge stored in repositories and papers into reusable, verified skills. The DisCo agent produces them in two forms: task-agnostic — yielding the AREX-Skill library of 5,000+ verified skills from 1,000 widely-used ML repositories, organized into 20 domains and 178 capability families — and task-oriented, generated for the task at hand. With a fixed GPT-5.5 backbone, test harness, and execution budget, skill-equipped agents score +134.3% on MLE-bench, +34.4% on PaperBench, +9.2% on FrontierCS, and +14.0% on PassNet.
Imported from the existing research archive. Migration preserves the recorded judgment and is not a fresh verification.
What this means for you
The skills direction (Claude Skills, MCP tool libraries) gets its largest quantified result: a verified skill library multiplies agent capability at fixed model and execution budget — the gains come from knowledge, not scale. The task-agnostic library recipe applies directly to internal tooling: distill your own repos into skills once, reuse across agents.
FRONTIER WATCH
Public artifacts and validated results are different.
The recorded fact is the public work and its artifacts. Author-reported outcomes do not establish independent reproduction. Next checks include usable code, task conditions and third-party validation.
Evidence & related material
Technical details & scores
| Arxiv | 2609.02749, announced in the Thu 3 Sep 2026 digest |
|---|---|
| Agent | DisCo agent; two distillation modes: task-agnostic and task-oriented |
| Library | AREX-Skill: 5,000+ verified skills from 1,000 widely-used ML repositories; 20 domains, 178 capability families |
| Results | fixed GPT-5.5 backbone + harness + execution budget: MLE-bench +134.3%, PaperBench +34.4%, FrontierCS +9.2%, PassNet +14.0% |
Technical impact 3
Engineering value 4
Adoption signal 2
Maturity 2
Verification cost 3
Risk 2