STUPID-2026-0014

Devin CI workflow caused 836-comment spam storm on single PR

3.0low
March 21, 2026Source checked

Why 3/10

836 bot comments notified every watcher of the repo; nothing was lost, the PR was abandoned.

Scope
1/4
A few files, one session, one credential, or a small cost
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

    Migrate to GitHub Container Registry and add version management

  2. Expected behavior

    Submit a clean migration PR that passes CI

  3. Actual behavior

    836 comments generated on a single PR. PR never merged. Evidence of CI/agent feedback loop where automated responses kept triggering each other.

  4. Damage

    836 comments of noise on a single PR. GitHub notification spam for all repo watchers. PR abandoned.

Full write-up

A Devin PR to migrate a project to GitHub Container Registry on arnaudlh/rover generated 836 comments — overwhelmingly automated CI feedback loops and Devin auto-responses. The PR was never merged. The volume of comments suggests the agent was caught in a CI feedback loop where each fix triggered new CI comments, which triggered new Devin responses, which triggered new fixes.

Classification

Agent
Devin
Failure mode
Infinite Loop
Root cause
Tool Misuse
Domain
Infra
Language
Yaml
MIT AI Risk Repository (our mapping)
7.3 Lack of capability or robustness taxonomy
Source
Github Pr

Related incidents

9.0
Clinejection: an AI issue-triage workflow enabled arbitrary code execution on the CI runner
2.0
Devin repeatedly submitted identical docs PRs that kept getting rejected
2.0
Devin cross-platform CI added 8-comment review cycle without landing
8.0
Malicious cloned repository triggered code execution in Cursor on Windows
Discuss on Hacker News
Cite this record
StupidLLM. "STUPID-2026-0014: Devin CI workflow caused 836-comment spam storm on single PR." AI Agent Incident Database, 2026. https://www.stupidllm.com/incident/STUPID-2026-0014/ (accessed 2026-10-29). CC BY 4.0.

← STUPID-2026-0013By IDSTUPID-2026-0015 →

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.