Created a PR to Migrate

openhands cloud

to use V1 API endpoints https://github.com/OpenHands/OpenHands-CLI/pull/643 Can anyone take a look ?

been out for a bit - what i miss...

I’m on OpenHands CLI 1.14.0 and the conversation stalls at the end. • It displays “Working (368s • ESC: pause)” and it stays there (the seconds counter keeps increasing). • The agent has responded and the laminar traces show FinishAction but the UI is unresponsive, and I can’t continue the conversation. • Pressing ESC doesn’t help, it displays “Pausing conversation this may take a few seconds” but it stays there. Any tips?

FYI, a community member has submitted a PR for migrating

openhands cloud

to V1 https://github.com/OpenHands/OpenHands-CLI/pull/643

PR to get the PR Template to align with OH/OH Ready for review and merge if no feedback ty https://github.com/OpenHands/OpenHands-CLI/pull/653

PR to fix AWS Bedrock auth for CLI when using IAM credentials (no API key needed). Fixes the case where

--override-with-envs

would bail on missing

LLM_API_KEY

for Bedrock/SageMaker models. Ready for review + merge since I can't merge after approval https://github.com/OpenHands/OpenHands-CLI/pull/612

Does anyone know the general intention of using

--mount-cwd

? I assumed from reading it would let me use a local project in the gui without needing to start it in the gui or add it as externally hosted git repo. After running through it a few times and asking OH about it that isn't the case. The mounted folder goes into the /workspace and then the bash_events, conversations, and project folder are added and used for any conversations. Feels like I'm missing something.

Does anyone have any interest in agentsview which is a local front end for viewing agent sessions. Its a nice way to view /search all your sessions in one place, whether its openhands or another agent like codex. I noticed they didn't have support for openhands CLI, but I can submit one if there is interest. - https://github.com/wesm/agentsview (it would be nice to get someone validating that everything is running as it should before I submit a PR)

I have a few quick questions: 1.

make test

seems to write to

~/.openhands

instead of staying isolated. Is that expected? Also, if I set

OPENHANDS_PERSISTENCE_DIR

to work around that, one test fails (

test_work_dir_and_persistence_dir_are_different

). 2. I was thinking about opening a PR to add devcontainer support to the CLI repo, similar to the GUI repo. But I'm not 100% sure it's a good idea, since it might push new contributors toward a clunkier setup (testing/using the CLI from the VS Code terminal, I'm not sure it works well there). Would appreciate a yes/no on this one. 3. I also wanted to add podman support in

openhands serve

(also handling rootful/rootless+selinux cases for both docker and podman, where possible). Would it be ok to add it as a separate command, something like

openhands serve-new

? I'm not sure how good it is going to be, and I don't want to risk breaking

serve

.

Have a PR to [CLI]: Add rich diff viewer inside the TUI Could anyone please take a look?

I had someone that had trouble connecting their OpenAI- proxy (Bifrost) with the CLI - does it make sense to add more information in the docs on connecting with the CLI. Something along the lines of:

# This already works today:
export LLM_BASE_URL="<http://localhost:8080/v1>"
export LLM_MODEL="bedrock/anthropic.claude-sonnet-4-6"
export LLM_API_KEY="dummy"
openhands --override-with-envs

I didn't see anything in the docs, but maybe I missed it.

Been blocked from using CLI for a while due to this, when prompted to update the settings I can't hit enter, can't Ctrl-C out, have to kill the process.

OpenHands CLI doesn't clean up after itself on exit. I often have to reset the terminal after running it or use CTRL+J to respond to y/n prompts if I have forgotten to reset. Been meaning to file an issue on that.

Created a PR to [CLI]: Add rich diff viewer inside the TUI Can anyone take a look ?

Are we tracking CLI startup time? I feel like it's significantly degraded

like ~20s to start up

Have a PR to Support /skill lifecycle commands for installed AgentSkills Can anyone please take a look?

I have a few PR's to improve keyboard navigation if anyone has some time to review. Thanks. Allow Ctrl-C/Ctrl-Q/ESC to work everywhere including double Ctrl-C for fast exit: https://github.com/OpenHands/OpenHands-CLI/pull/686 Fix tab navigation in settings modal: https://github.com/OpenHands/OpenHands-CLI/pull/687 Convert History panel to ListView so arrow keys work with it: https://github.com/OpenHands/OpenHands-CLI/pull/688

Sub-agents appear to have stopped working in 1.15.0 release as part of switching away from deprecated Delegation tool (now get an error message with

no running event loop

) but Gabriel Werneck has raised a PR. Posting this here in hopes it can make it in before next release. https://github.com/OpenHands/OpenHands-CLI/pull/695

PR that clears max_input_tokens on mode and model changes in addition to base_url change (for example switch from gemini to gpt now resets to meaningful max tokens): https://github.com/OpenHands/OpenHands-CLI/pull/697 Running targeted model change tests (instead of using make to run all) was corrupting ~/.openhands/agent_settings.json so moved prevention from Makefile to test framework: https://github.com/OpenHands/OpenHands-CLI/pull/699

@quick-window-91705 Could you review this PR? https://github.com/OpenHands/OpenHands-CLI/pull/648 Thanks 🙂

Minor focus fix: Pre-existing feature moves the focus to the prompt input if user types a printable character but the character was discarded. https://github.com/OpenHands/OpenHands-CLI/pull/698

Hey @quick-window-91705, no rush to do it today, but apparently our CLI release has pinned an older version of the SDK that doesn’t invoke skills in the standard way, so if it’s not too hard it might be nice to get a new release (or even just a patch release that bumps the SDK version). (I CCed you on a related email)

Went to report that cli login is broken for me, looks like @limited-student-10747 already reported it so added my video to that. https://github.com/OpenHands/OpenHands-CLI/issues/705

I'm looking for feedback on 1 specific part in this PR that adds prompt history support. It creates a ~/.openhands/projects to store things that make sense to store for a specific project directory instead of storing a single global version. I'm proposing that storing history per project directory is better than a single huge global history. Longer term we could store things like auto-approve mode at a project directory level as well. Thoughts? https://github.com/OpenHands/OpenHands-CLI/pull/715

🙋 I realize we were missing a guide for contributing and development in the repo. I added that and a reference to these two new files in the readme. My PR is ready to review https://github.com/OpenHands/OpenHands-CLI/pull/721 One big note: I have not contributed to the openhands-cli so IDK if the development.md is accurate. I would greatly appreciate if someone familiar to testing changes to the CLI can review that document. all-hands-bot approved it for merging but thinky spin idk if I trust it closing PR based on feedback ty

A continuation of @mammoth-pillow-18568 PR to replay conversation events when switching conversations (now with ability to load more if limit is hit): https://github.com/OpenHands/OpenHands-CLI/pull/734

Hi all do we have a native windows installer ready

any issues to pick up?

#615 ready for review, fixed the plan mode escaping into edits bug 👀 github.com/OpenHands/OpenHands-CLI/pull/615