Track target: UiPath AgentHack 2026, Track 3 - UiPath Test Cloud
Agentic Release Risk Copilot helps enterprise automation teams decide what to test before a release. It reads a requirements note, repository diff, test inventory, and failed-test log, then produces:
- Risk-ranked release areas
- Suggested regression scenarios
- Brittle selector warnings
- Remediation recommendations
- Human approval checkpoints for high-risk changes
The local prototype is deliberately dependency-free so it can run before UiPath Labs access arrives. Once Labs access is granted, UiPath Automation Cloud/Test Cloud becomes the orchestration and governance layer around this analysis flow.
In the final AgentHack version, UiPath coordinates the long-running release validation process:
- A UiPath workflow ingests change artifacts from GitHub, CI, or Test Cloud.
- A coded agent performs risk analysis and generates test recommendations.
- Test Cloud executes or schedules the selected test set.
- High-risk releases route to a human approval step before deployment.
- Results are logged as auditable release evidence.
python3 -m unittest discover -s tests
python3 src/release_risk_copilot/cli.py --output out/demo-report.json --markdown out/demo-report.mdThe command uses the sample data in data/ and writes a JSON report plus a Markdown demo report.
The report includes:
summary: total risk count and highest severityrisk_items: area-level risk findingstest_plan: recommended scenarios to runhuman_approval_queue: items requiring reviewer approvaluipath_next_steps: how to connect the prototype to UiPath Labs
data/ Synthetic sample inputs
docs/ Architecture and submission drafts
src/release_risk_copilot/ Dependency-free prototype
tests/ Unit tests
This is a working local prototype and documentation scaffold. The remaining external blocker is UiPath Labs access, which was requested through the official form and may take up to 3 business days.