STUPID-2026-0013

Devin attempted to build entire Figma clone from scratch — 3 rejected attempts

2.0low
March 21, 2026Source checkedRepro reported

Why 2/10

Three failed PRs burned compute and review time; no code landed.

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

    Build Vigma design tool

  2. Expected behavior

    Start with a minimal viable implementation, iterate based on feedback

  3. Actual behavior

    Three separate full-rewrite attempts, each massive in scope, each rejected. No incremental approach. Tried different UI frameworks each time but same overambitious scope.

  4. Damage

    Three failed PRs consuming significant compute/API resources. No usable code delivered.

Full write-up

Devin submitted 3 separate PRs to andrewgcodes/vigma, each attempting to build a full Figma-like design tool from scratch. PR #4 ("Full-featured Vigma design editor with Apple/Stripe style UI"), PR #5 ("canvas-based design tool with Fabric.js"), and PR #6 ("Sleek minimalist design tool"). All were closed. The agent kept attempting the same massive scope task with different frameworks (Fabric.js 6 in one, different versions in others) without reducing scope or understanding why previous attempts failed.

Classification

Agent
Devin
Failure mode
Scope Explosion
Domain
Frontend
Language
Javascript
MIT AI Risk Repository (our mapping)
7.3 Lack of capability or robustness taxonomy
Source
Github Pr

Related incidents

Cite this record
StupidLLM. "STUPID-2026-0013: Devin attempted to build entire Figma clone from scratch — 3 rejected attempts." AI Agent Incident Database, 2026. https://www.stupidllm.com/incident/STUPID-2026-0013/ (accessed 2026-10-29). CC BY 4.0.

← STUPID-2026-0012By IDSTUPID-2026-0014 →

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.