{
  "id": 14943639,
  "title": "Your AI coding agents are reading conflicting instructions",
  "url": "https://urgent.news/2026/10/16/your-ai-coding-agents-are-reading-conflicting-instructions",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-16T11:45:00.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/jbrenens/your-ai-coding-agents-are-reading-conflicting-instructions-1ie4"
  },
  "original_language": "en",
  "account": "Running Claude Code, Codex and Cursor on the same repository for a week resulted in multiple conflicting instruction files, including a CLAUDE.md, AGENTS.md, .cursor/rules folder and a README. Some files instructed to use pnpm, while others allowed npm. Other files stated not to touch migrations, while others provided a migration workflow. Agents did not identify these conflicts; they simply followed the last read rule or the closest relevant rule. This led to behavior that changed based on which tool was opened. To address this, several steps were recommended. First, locate all instruction files and list the contradictions. Next, declare a precedence order with the live human message, safety rules, nearest scoped file, canonical root RULES.md, agent-specific pointer files, and global home-dir files as the hierarchy. If two rules conflict, stop and ask for clarification instead of making an arbitrary choice. The final step is to consolidate the rules into a single RULES.md file with brief pointer notes in CLAUDE.md and AGENTS.md. This consolidates all rules into one location and prevents agents from drifting apart. A conflict log should also be maintained for audit purposes.",
  "summary": "Run Claude Code, Codex and Cursor on the same repo for a week and you'll end up with a CLAUDE.md , an AGENTS.md , a .cursor/rules folder and a README that each say something slightly different. One says \"use pnpm\", another says \"npm is fine\". One says \"never touch migrations\", another has a migration workflow. Agents don't flag these conflicts. They just follow whichever rule they read last, or…",
  "key_points": [
    "Multiple AI coding agents reading conflicting instructions caused issues.",
    "No single agent identified these contradictions, following last read rule."
  ],
  "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."
}