# Evaluation: 2026-07-05-devstral-24b-self-repair-001

Verdict: failed

Functional status: broken

Scores:

- Interest: 1/5
- Execution: 1/5
- Taste: 1/5

## Summary

self_repair artifact did not pass the local HTML smoke test.

## Findings

- Artifact references external URLs or remote fetches: https://unpkg.com/leaflet@1.7.1/dist/leaflet.js
- Runtime smoke error during DOMContentLoaded 1: ReferenceError: L is not defined

## Claim Accuracy

Not accurate enough under automated smoke testing; the generated artifact still has runtime, self-containment, or interaction issues.

## Notes

Automated evaluation from `scripts/smoke_html.py`. This is a functional smoke score, not a full manual artifact review.

Source run: `runs/2026-07-05-devstral-24b-001`

## Artifact

Entry point: `runs/2026-07-05-devstral-24b-self-repair-001/artifact/index.html`