{
  "id": 17222639,
  "title": "I tried SKILL.state in my coding agents and stopped after four days",
  "url": "https://urgent.news/2026/10/30/i-tried-skill-state-in-my-coding-agents-and-stopped-after-four-days",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-10-30T12:20:40.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/anton_shubin_f26b3cb5a572/i-tried-skillstate-in-my-coding-agents-and-stopped-after-four-days-151b"
  },
  "original_language": "en",
  "account": "SKILL.state is a runtime design for long agent tasks that replaces the growing chat history with an explicit, mutable execution state. At each step, the model receives only the skill, the current state, and the latest observation. The paper suggests improved accuracy and lower token usage. The author tested SKILL.state on real tasks by running it alongside a normal agent for four tasks. Despite the briefs being good and no agent corrections needed, the author encountered several issues that required manual fixes, such as schema rejection, invented timestamps, and broken state folder checks. Additionally, the introduction of a state file added more handoffs, leading to more potential errors. The author stopped after four tasks and then switched to a model with a smaller context window. After moving to Opus 5.5, the problem that SKILL.state aimed to solve became less significant, and the author found that a fresh agent for each step with a self-contained brief already provided the benefits without the overhead.",
  "summary": "A friend sent me the SKILL.state paper and later asked how it worked out. This is the answer: what I built from it, the four tasks I ran, and why I stopped. It is for anyone who runs coding agents on long tasks and wonders whether an explicit state file is worth it. TL;DR SKILL.state swaps an agent's growing chat history for a state file and a fresh step each time. I ran my own version of it on…",
  "key_points": [
    "SKILL.state replaces chat history with explicit execution state",
    "Author tested on four tasks, encountered manual fixes",
    "Switched to Opus 5.5 model, benefits lessened"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}