Portfolio: 43.9% weighted coverage · 15/29 scenarios · 2/3 projects stable
Agent telemetry
Autonomy Trends
Promoted scenario evidence and recent committed runs extend these charts over time.
Portfolio coverage
43.9%
Repair effectiveness
0%
Flakiness rate
0%
Live CI telemetry
30-Day QA Health
Waiting for the first published agent run.
| Application | Regression reliability | Candidate success | Timeouts | Latest result |
|---|---|---|---|---|
| Waiting for published run data. | ||||
Recent Activity
Latest published runs
No live runs published yet
The scheduled agent workflow will populate this section automatically.
Current coverage
Web Applications
Permanent regression suites and agent-generated candidates are grouped by application.
How it works
Agent Pipeline
- ExplorePlaywright opens the target page and the model chooses safe next actions.
- GenerateThe model turns the exploration trace into Playwright tests.
- ExecuteThe generated spec runs in Chromium with traces and screenshots on failure.
- RepairFailures are sent back to the model for selector and assertion repair.
- ReportRuns produce dashboard data and Playwright reports for review.
Next steps
Portfolio Roadmap
Live Report Publishing
Publish the latest dashboard and report artifacts to a stable URL.
Candidate Promotion
Review validated agent candidates and promote useful coverage into the permanent regression suite.
Specialized Agents
Split exploration, repair, accessibility, and reporting into focused roles when complexity grows.