Trendoday

From Repositories to Training Environments: Scaling Long-Horizon Coding Agents with Environment Evolution Loop

Training coding agents for long-horizon tasks requires environments with verifiable feedback. Existing pipelines often follow a linear process that ends with validation, establishing usability but pro

  • Upvotes9
–Rank now
#8Peak rank
2Days on list
201Score · 7d
Nov 18, 2026First seen
Rank · 7 days

About this paper

RepoFactory converts real repositories into verifiable coding environments and optimizes them via an Environment Evolution Loop using proxy feedback, enabling trained coding agents to achieve strong gains on long-horizon coding and agentic benchmarks.

Training coding agents for long-horizon tasks requires environments with verifiable feedback. Existing pipelines often follow a linear process that ends with validation, establishing usability but providing little guidance on improving training value. Recursive self-improvement (RSI) often requires model rollouts or model training followed by downstream evaluation, with costs and feedback delays amplified on long-horizon tasks. How can we use timely, reliable proxy signals to improve training value and synthesize environments at scale? We present RepoFactory, an automated pipeline that converts real repositories into verifiable environments where agents build complete projects from scratch…

coding agentslong-horizon tasksverifiable feedbackrecursive self-improvementmodel rolloutsdownstream evaluationproxy signalsverifiable environmentslanguage specificationEnvironment Evolution Loop
Upvotes
9
Authors
8
Published
Nov 17, 2026

Summary and abstract excerpt via Hugging Face Papers ↗ · arXiv.

Trending history

From Repositories to Training Environments: Scaling Long-Horizon Coding Agents with Environment Evolution Loop was first seen on Hugging Face · Daily papers on Nov 18, 2026. Since then it has appeared in 12 checks on 2 different days. Its best position was #8 of 30, reached on Nov 19, 07:29 UTC. It dropped off the list after Nov 19, 07:29 UTC.

Rank history

Position on Hugging Face · Daily papers at every check (lower is better).

#1#11#20#30 09:2814:5820:2901:5907:29 Nov 18, 09:28 UTC · #20Nov 18, 11:28 UTC · #21Nov 18, 13:28 UTC · #19Nov 18, 15:28 UTC · #17Nov 18, 17:28 UTC · #15Nov 18, 19:28 UTC · #15Nov 18, 21:28 UTC · #14Nov 18, 23:28 UTC · #14Nov 19, 01:28 UTC · #14Nov 19, 03:28 UTC · #14Nov 19, 05:28 UTC · #10Nov 19, 07:29 UTC · #8

Last 14 days

Best rank per day (UTC).

Nov 9today

Metric history

Platform metric as reported by Hugging Face.

04812 09:2814:5820:2901:5907:29 Nov 18, 09:28 UTC · 9Nov 18, 11:28 UTC · 9Nov 18, 13:28 UTC · 8Nov 18, 15:28 UTC · 8Nov 18, 17:28 UTC · 10Nov 18, 19:28 UTC · 10Nov 18, 21:28 UTC · 11Nov 18, 23:28 UTC · 11Nov 19, 01:28 UTC · 12Nov 19, 03:28 UTC · 12Nov 19, 05:28 UTC · 12Nov 19, 07:29 UTC · 12

Questions

Is From Repositories to Training Environments: Scaling Long-Horizon Coding Agents with Environment Evolution Loop trending on HF Daily Papers right now?

Not at the moment. It was last on the list on Nov 19, 07:29 UTC, after 2 days of trending.

What is the highest rank From Repositories to Training Environments: Scaling Long-Horizon Coding Agents with Environment Evolution Loop reached?

#8 on Hugging Face · Daily papers, on Nov 19, 07:29 UTC. Trendoday checks this list every 120 minutes.

How long has From Repositories to Training Environments: Scaling Long-Horizon Coding Agents with Environment Evolution Loop been trending?

It was first seen on Nov 18, 2026 and has been on the list on 2 days (12 checks) since.