[deep-report] Visual Regression Checker lacks a timeout: 2 of last 3 runs hung 1.6-2.0 hours before failing

[github-actions]

Problem

The Visual Regression Checker workflow has no effective timeout: 2 of its last 3 runs hung for 1.6-2.0 hours before eventually failing, versus a clean ~4-minute driver_exit on the third (healthy) run. A workflow that can silently occupy a runner for ~2 hours before failing wastes CI capacity and delays feedback far longer than the failure itself warrants.

Evidence

• Source: Audit Workflows discussion #56739, finding 5 — reports the 1.6-2.0 hour hang-before-fail pattern across 2 of the last 3 runs, contrasted with a ~4 minute clean exit on the healthy run.

Suggested action

Add an explicit timeout-minutes (or equivalent engine-level timeout) to the Visual Regression Checker workflow, set well above the healthy-run duration (e.g. 15-20 minutes) but far below the observed 1.6-2.0 hour hangs, so failures surface promptly instead of consuming a runner for hours.

Acceptance criteria

• Visual Regression Checker workflow has an explicit timeout configured.
• A simulated/forced hang in this workflow is terminated by the timeout well under an hour.

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• api.anthropic.com

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "api.anthropic.com"

See 网络 Configuration for more information.

Generated by 🔬 Deep Report · claude · agent · 258.7 AIC · ⌖ 8.26 AIC · ⊞ 12.4K · ◷

• expires on Aug 30, 2026, 10:49 PM UTC-08:00