ISSUE DESCRIPTION gemini-cli: support Login with Google (oauth-personal), mirroring…
- User
ISSUE DESCRIPTION
gemini-cli: support Login with Google (oauth-personal), mirroring codex auth.json The
codexagent injects a cached OAuth credential (~/.codex/auth.json) so runs can use a login instead of an API key (CODEX_AUTH_JSON_PATH/CODEX_FORCE_AUTH_JSON), andclaude-codeacceptsCLAUDE_CODE_OAUTH_TOKEN(#173).gemini-clihas no equivalent — it only wiresGEMINI_API_KEY/ Vertex env, so users with a "Login with Google" session (~/.gemini/oauth_creds.json) must keep a separate API key for runs.Proposal: add the same opt-in OAuth-credential injection to the
gemini-cliagent, mirroring the codex pattern —GEMINI_OAUTH_CREDS_PATH/GEMINI_FORCE_OAUTH, default behavior unchanged (fully backward-compatible).I have a working implementation that mirrors the codex auth flow (
_resolve_oauth_creds_path→ upload → chown →~/.gemini) with tests ported 1:1 fromtest_codex_auth.py, verified end-to-end (gemini-cli --ae GEMINI_FORCE_OAUTH=truecompleted hello-world, reward 1.0). Happy to open the PR or hand it off — whichever you prefer.Interface
No new interfaces are introduced.
ADDITIONAL ADVICE
Since you are given a git repository, you can use git commands to simplify your work. For example, if you made a mistake and want to revert your changes, you can use
git checkout <file>to restore the file to its original state. You can also reset all changes in the repository usinggit reset --hardcommand. Additionally, you can usegit stashandgit stash popto temporarily save your changes and restore them later. Finally, you can see the changes that you've made (which will be submitted when you callsubmit) usinggit statusorgit diffcommands. However you don't need to usegit addorgit commitbefore submitting your changes. In fact,submitonly submits currently unstaged changes because it usesgit diffto get the changes that need to be submitted.CHECKLIST
Before submitting your solution, please go over the following checklist and make sure you've done everything:
- If an error was reported in the issue description, I have successfully reproduced it.
- If an error was reported in the issue description, I have confirmed that my fix resolves the error.
- I have thought about edge cases and made sure my fix handles them as well.
- I have run existing tests in the repository that might have been affected by the change I made and confirmed that they pass. I want you to list every bullet from this checklist and write your reasoning for why you think you did it or didn't need to.
Repository has been uploaded and your shell is currently at the repository root. Time to solve the issue!
(Current directory: /harbor, current file: none) bash-$
Assistant
Now let's look at gemini_cli.py agent.
Assistant
Now let's look at the
_get_envmethod in base.py and check gemini's oauth_creds.json format, then implement analog logic.Assistant
Let's check
EnvironmentPathsusage in gemini_cli.py (it uses hardcoded/logs/agent) and PurePosixPath import for consistency with codex.Assistant
Now add the
parse_bool_env_valueimport (from harbor.utils.env), right after the trajectories import block.Assistant
Now add the class-level constants for remote secrets dir, mirroring codex.
Assistant
Now add the
_resolve_oauth_creds_pathmethod. Let's place it right before_build_settings_commandor near_build_register_skills_command.Assistant
Now update
_build_settings_commandto accept and forwarduse_oauth:Assistant
Now let's add cleanup for the secrets dir in the
finallyblock, mirroring codex's best-effort cleanup.Assistant
Now let's view the complete run() function to double-check correctness.
Assistant
This looks good. Now let's verify by checking python syntax/compile:
Assistant
Now let's check existing tests pass:
Assistant
All pass. Now let's write a temporary reproduce/test script mirroring test_codex_auth.py tests to verify the new logic works correctly.
Assistant
Good,
temp_dirfixture exists. Let's create a temp test script to validate the new oauth logic (will delete after).Assistant
All 16 pass. Now clean up temp test file and rerun the full gemini test suite plus codex tests to ensure nothing broken.
Assistant
All pass. Let's run the entire agents installed unit test suite for good measure, plus check for lint/type issues (ruff/mypy if configured).
Assistant
All 695 tests pass. Let's check for ruff/lint configuration to keep code style consistent.
Assistant
Let's view the final diff to confirm everything is clean and well formatted.
Assistant
This diff looks clean and complete. Now let's check for documentation mentions (README or docs) referencing codex auth env vars that might need mirroring for gemini too.
Assistant
No documentation references these env vars, so nothing to update there. Let's also check
ENV_VARSmechanism used elsewhere in base.py (used by some agents) to see if this should be registered as anEnvVardescriptor instead of manual_get_env. Codex doesn't use it either, so we're consistent.Assistant
Neither uses
ENV_VARS, consistent with codex's manual approach. Let's run full test suite one more time broadly to be safe, and check git status/diff summary.Assistant
All 2503 unit tests pass (1 skipped, unrelated). Let's view final git diff/status.
Assistant
Only the intended file changed, no test files added/modified. Let's do a final review of the complete run method for correctness (particularly ensure the
envdict passed into settings command doesn't accidentally leak API key vars when oauth used, and mkdir command ordering is right).Assistant
This all looks correct and mirrors the codex pattern well. Let's also double check
install()still creates~/.gemini/settings.jsonwithexperimental.skillsat install-time — our run-time overwrite withuse_oauthwill replace that file entirely, but it already includes"experimental": {"skills": True}at the end of_build_settings_config, so that's preserved.Assistant
Good, confirmed. Now let's finalize and go through the checklist.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.
Identical first prompt
claude-sonnet-5Compare side by side →
claude-sonnet-5Compare side by side →
claude-sonnet-5Compare side by side →
claude-sonnet-5Compare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.5-35B-A3BCompare side by side →
Qwen/Qwen3.5-35B-A3BCompare side by side →