Use one agent lifecycle for LangChain, LangGraph and custom deployment policies
Added to DistroCool on 2026-10-22 · Discovery date, not a release announcement
Google Agents CLI 1.8.0 adds an experimental extension system and a first-party LangChain template. It can scaffold, run, deploy and evaluate LangChain or LangGraph agents through the same lifecycle used for ADK, while custom extensions can wrap commands such as deploy with organization-specific checks.
Why it matters
A team can keep a common deployment, evaluation and observability path without forcing every agent onto one framework or maintaining a forked CLI. Extensions are pinned to exact Git commits in a lockfile, which helps make local and CI behavior reproducible. The extension system is experimental and may introduce breaking changes, so teams should pin the Agents CLI version; Google Cloud billing and the documented Python, uv and gcloud setup are required.
Try this next.
In a sandbox Google Cloud project, pin Agents CLI 1.8.0 and scaffold the official LangChain template. Run its sample evaluation before deployment, inspect which commands the extension overrides, and commit the extension lockfile. For a policy proof of concept, wrap deploy with a harmless local check before attempting an SBOM gate. DistroCool has not independently tested this example.
Open the practical workflow →Source: Google Developers Blog — October 21, 2026 ↗