Hi playwright-cli team,

The release of the Playwright CLI is great and thanks for contribution to the community. I’ve realized that there are already many CLIs available (nearly 30 or more), and it’s not easy for either humans or agents to remember them or use them correctly on the first try (even for agents with Skills definition). It’s especially difficult for human users to recall commands or constantly check the README.md or --help.

We are developing an agent/human CLI hints and completion tool to address this. With it, when users want to use Playwright to open a web page, they can simply type “play” or “open,” and the appropriate CLI commands will be retrieved and completed interactively for humans, while providing terminal hints for agents.

Example

agtm run play # Uses hints to resolve to the Playwright CLI
agtm run microsoft/playwright-cli open  ## "goto" is hinted as a Playwright command (navigate to a URL)
# agent mode Non Interactive Hints
agtm run microsoft/playwright-cli open --mode agent ## agent mode return hints for similar commands of keyword "open"

Agtm Cli Hint Exampls Snapshot

agtm run play: Hints the Skills

agtm run microsoft/playwright-cli open: Hints the Command of "open"

agtm run play: Complete and Run if user prompt their url

Docs & Usage of agtm

npm install -g @aiagenta2z/agtm
npx agtm setup --hint

GitHub
https://github.com/aiagenta2z/agtm,
https://github.com/aiagenta2z/onekey-gateway
Documentation

And we would like to share the new tool to the community and just add a small part to the README.md with "summary tag" below the Commands section.

Would you please help review the PR and where would you recommend appropriate to make the contribution?

Best Regards,
AI Agent AZZ x DeepNLP team