- Who it helps
- Developers changing an agent runtime or porting an existing workflow.
- Availability
- The change is documented in @statelyai/agent 2.0.0-alpha.30. It remains a prerelease; integration and model-service costs depend on the application.
Compare the result and effects
The helper runs supplied scenarios through baseline and candidate implementations. Its labels distinguish unchanged behavior, intended improvement and regression using outcomes and effects. It does not require identical tool-call order.
Keep context with the run
The same release adds host-owned JSON metadata for policy versions, model identities or scenario information. Resuming preserves the initial stamp. A useful scenario should check external effects as well as a plausible final response.
What you can do
- Review the release notes and choose a known failure scenario before evaluating a port. Open reference ↗
Where the evidence stops
Release notes were read; helper code was not executed. Classification is only as useful as the scenarios, rules and expected outcomes supplied.
Sources read. Checked October 23, 2026. The alpha.30 release notes were read. The comparison helper and metadata behavior were not tested.
Sources & further reading
GitHub · implementation
Stately Agent alpha.30 release notes ↗
Read October 22 compareRuns and run-metadata changes; package remains a prerelease.
Stately · page checked 2026-10-23 · Published 2026-10-22
Coverage note: The alpha.30 release notes were read. The comparison helper and metadata behavior were not tested. Original video and X coverage remain incomplete in this bounded sample.
The record over time (1)
Dated source checks and published explanations. An observation date can differ from the original release date.
- · Added to Current
Stately’s October 22 prerelease adds compareRuns, a helper for comparing a candidate agent implementation with the version it replaces.