- Test closure variable detection (helpers, constants, imports)
- Test nested dependency resolution and ordering
- Test class definitions and decorator removal
- Test import style preservation
- Test caching mechanism
- Implement AST-based dependency analysis with closure support
- Automatically detect and include helper functions, imports, and constants
- Handle nested dependencies recursively
- Add caching with function object as key
The TypeScript CloudComputer was using a stale hardcoded domain
(containers.cloud.trycua.com). Now fetches the host from the /v1/vms
API like the Python implementation, with fallback to sandbox.cua.ai.
Fixed the Lume macOS install script failing with 404 errors when trying
to download release files. The script was using /releases/latest which
resolved to the most recent cua CLI release instead of the latest lume
release.
Added documentation for accessing Claude models through Azure AI Foundry
as an alternative to direct Anthropic API access. Users can now use the
same anthropic/ prefix with Azure-hosted endpoints by configuring
ANTHROPIC_API_BASE.
Split the final metadata-action step into PR, main, and semver-specific blocks so each
event only generates the appropriate tags. This prevents PR runs from pushing
`latest` or semver tags, ensuring the publish job creates multi-arch manifests
only for the tags intended for that event.
Updated the manifest creation and inspection steps to loop over metadata-action's `.tags[]`
directly and use the full tag references as-is. The previous version attempted to prefix
tags with the image name, which produced invalid references (e.g. image:image:tag). Using
the full tags emitted by metadata-action ensures correct manifest creation for PR tags,
SHA tags, semver tags, and latest.
Refactored the reusable Docker publish workflow to generate proper multi-architecture images.
Each platform build now pushes by digest and uploads its digest as an artifact, and a final job
assembles the multi-arch manifest from those digests.
Updated all Python code examples in the quickstart and agent sections to use explicit async main functions and asyncio.run(). This improves clarity for asynchronous usage, ensures proper connection and disconnection of resources, and aligns with best practices for async Python code.
