STUPID-2026-0012

Devin repeatedly submitted identical docs PRs that kept getting rejected

2.0low
March 21, 2026Source checkedRepro reported

Why 2/10

Nine rejected PRs cost maintainer time; nothing was merged or damaged.

Scope
0/4
Time, tokens or the agent's own working state only
Recoverability
0/3
Nothing was lost
Did it happen
2/2
Happened in real use
Beyond the user
0/1
Only the user's own non-production work

The score is the sum of these four factors. How each factor is scored · scored 2026-10-29

  1. Instruction given

    Fix documentation drift in datastack repos

  2. Expected behavior

    Submit a fix, if rejected, understand why and adjust approach

  3. Actual behavior

    Submitted 5+ identical PRs to the same repo, each closed without merge. No variation in approach between attempts. Same title, same content, same result.

  4. Damage

    Spam-like behavior. 9 rejected PRs across 2 repos for the same issue. Wasted maintainer time reviewing duplicate submissions.

Full write-up

Devin submitted 5 nearly identical PRs to hailbee/datastack-docs-drift-demo, each titled "fix: update docs to match current API behavior." Each was closed without merge, but Devin kept submitting the same fix. This pattern repeated across turnup4ethan2/datastack-demo-api with 4 identical "docs: auto-update API reference" PRs. The agent showed no ability to learn from rejection.

Classification

Agent
Devin
Failure mode
Infinite Loop
Root cause
Pattern Matching Error
Domain
Other
Language
Markdown
MIT AI Risk Repository (our mapping)
7.3 Lack of capability or robustness taxonomy
Source
Github Pr

Related incidents

Cite this record
StupidLLM. "STUPID-2026-0012: Devin repeatedly submitted identical docs PRs that kept getting rejected." AI Agent Incident Database, 2026. https://www.stupidllm.com/incident/STUPID-2026-0012/ (accessed 2026-10-29). CC BY 4.0.

← STUPID-2026-0011By IDSTUPID-2026-0013 →

Get told when an agent breaks something

We document AI agent failures daily, severity-scored against a published scale. Each Monday, if any scored 7 or above that week — deleted data, leaked secrets, broken production — you get one email with the sources. When nothing did, you get nothing.

This database is callable over MCP — query it from inside your agent.