# anyone else getting “intent conflicts” when two coding agents run in parallel, not just merge conflicts? — retained record

> Retained reddit record from reddit:ClaudeAI.

Canonical: https://highsignal.app/data/records/380c184d95b52625

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ClaudeAI)  published 2026-09-25 01:03 ingested 2026-09-26 02:35

[original source ↗](https://www.reddit.com/r/ClaudeAI/comments/1wpixl9/anyone_else_getting_intent_conflicts_when_two/)

## retained text



```
i’ve been spinning two (sometimes three) agents on the same repo and the git side is mostly fine with worktrees. what keeps biting me is when both think they own the same job — one wants to refactor the helper, the other wants to delete it and inline, and both look “correct” until you read the intents next to each other.

curious if anyone’s doing anything before the write: like making agents declare scope / intent first and failing when those clash, vs just reviewing diffs after. worktrees solve stomping files; they don’t solve two confident plans that disagree.

war stories > tool pitches. what actually reduced the “both were right, both broke the design” moments for you?
```



document fetched 2026-09-26 00:23 · canonical https://reddit.com/r/ClaudeAI/comments/1wpixl9/anyone_else_getting_intent_conflicts_when_two
