google/agents-cli
Frameworks & Eval · tested 2026-07-21 · re-test due 2026-10-19 · by the Hlido desk, not the vendor
In short: Official Google tooling that makes any coding assistant better at scaffolding, evaluating and deploying ADK agents on Google Cloud — a force multiplier, not an agent itself.
3 PASS · 2 FAIL of 5 public-surface claims
Quick answer
google/agents-cli scores 80/100 (STEADY) on Hlido’s independent, hands-on test (reviewed 2026-07-21). STEADY (80) because it's genuinely useful first-party infrastructure with a clear, coherent purpose (the ADK-agent create-to-production lifecycle) and the trust that comes with official Google maintenance. Pricing was not findable on the public surface when tested.
This is a first-party Google project (under the `google` GitHub org, documented on the Google Developers Blog), and it's important to be clear about what it is: agents-cli is a CLI plus a skills package for building, evaluating and deploying AI agents on Google Cloud — the agents themselves are built with Google's Agent Development Kit (ADK), and agents-cli handles everything around them: scaffolding, evaluation, deployment, and observability. It's explicitly a tool for coding agents, designed to make Claude Code, Codex, Antigravity or any other assistant fluent in the ADK-on-Google-Cloud lifecycle — while also working standalone from your terminal (`agents-cli scaffold/eval/deploy`). The strategic read is that Google is doing what platform owners do: reducing friction from create to production so its Agent Platform becomes the path of least resistance. The credit here is real — first-party maintenance, a coherent create-to-production story, and local development possible with just an AI Studio API key. The honest limits: value is tightly coupled to Google Cloud and ADK, so it's a lock-in-adjacent choice; there's no pricing on the tool surface because the cost lives in the cloud resources it deploys; and as a young first-party tool the independent evidence of real-world reliability is still thin.
Why STEADY
STEADY (80) because it's genuinely useful first-party infrastructure with a clear, coherent purpose (the ADK-agent create-to-production lifecycle) and the trust that comes with official Google maintenance. Not VITAL because it is deliberately ecosystem-bound (ADK + Google Cloud), which is a real portability trade-off, and because it's early enough that independent reliability evidence beyond Google's own docs is limited.
Public-surface checklist
- PASS Homepage loads (required)
- PASS Primary value prop (required)
- PASS Cta present (required)
- FAIL Pricing or access
- FAIL Evidence or demo
What we saw
1 screenshot captured by the Hlido engine during the reviewed run (run-73b729c1e19f21ff-google-github-io). Our own captures — not vendor marketing material.
What it does well
- Official Google tool — first-party maintenance and documentation via the Agent Platform
- Coherent create-to-production lifecycle: scaffold, evaluate, deploy, observe
- Works both as a skill layer for coding assistants AND standalone from the terminal
- Local development possible with just an AI Studio API key (no immediate Cloud spend to try)
- Composes with multiple assistants (Claude Code, Codex, Antigravity, etc.)
What it fails at
- Tightly coupled to Google's ADK and Google Cloud — portability/lock-in trade-off
- No standalone value for teams not building on ADK
- Cost is opaque at the tool level because it lives in the Cloud resources deployed
- Young project — independent reliability evidence beyond Google's own docs is thin
Best for
- Teams building agents with Google's ADK who want a paved deploy path
- Developers using a coding assistant who want it fluent in ADK-on-Google-Cloud
- Anyone standardising agent scaffolding, eval and observability on Google Cloud
Not recommended for
- Teams committed to a non-Google stack or avoiding cloud lock-in
- Projects not using the Agent Development Kit
- Buyers wanting a portable, vendor-neutral agent framework
Pricing & access
- Pricing findable on the public surfaceFAIL
Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-07-18.
Compared to
-
microsoft/autogen
google-cloud-paved-path
AutoGen is a vendor-neutral multi-agent framework you assemble yourself; agents-cli is Google's opinionated lifecycle tooling around its own ADK. Choose agents-cli for a paved Google-Cloud deploy path, AutoGen for framework portability.
-
CrewAI
tooling-vs-framework
CrewAI is a build-anywhere agent-orchestration framework; agents-cli is not a framework at all but the scaffold/eval/deploy tooling around ADK. They sit at different layers — pair agents-cli with ADK if you're on Google Cloud, use CrewAI if you want stack independence.
Agent relevance
CLI SDK Behavioral-testable
Agentic-Commerce Readiness 52/100 · INTEGRABLE
Independent readiness for agent delegation & transaction. How it’s scored · check live
High for the ADK-on-Google-Cloud niche. It is itself agent-infrastructure: a CLI + skills that coding agents consume to scaffold/eval/deploy other agents. Value is conditional on building with Google's ADK.
Agent-friendly score: 7/10
Score over time
The longitudinal record — every point is the score as published on that date. Raw series.
Evidence
- Official Google project: CLI + skills for ADK agents on Google Cloud — source (2026-07-21) verified
- Scaffold, evaluate, deploy, observe lifecycle; works standalone or with coding assistants — source (2026-07-21) verified
- Covered by Google Developers Blog (Agent Platform) — source (2026-07-21) verified
- Local dev with AI Studio API key; requires Python 3.11+, uv, Node.js — source (2026-07-21) verified
scorecard.json · transparency passport · registry · methodology
Embed this trust badge
Markdown
[](https://hlido.eu/check/?agent=google-agents-cli)HTML
<a href="https://hlido.eu/check/?agent=google-agents-cli"><img src="https://hlido.eu/badge/google-agents-cli.svg" alt="Hlido trust score"></a>