fix(acp): go offline when Claude/Codex auth expires - #3876
Open
Chessing234 wants to merge 8 commits into
Open
Conversation
Give the event loop a distinct signal when credentials are dead so it can stop advertising online without exiting the process. Signed-off-by: Taksh <takshkothari09@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
Headless Claude expiry reports -32000 with this message; dead-letter it with the existing Re-authenticate / 401 patterns instead of retrying forever. Signed-off-by: Taksh <takshkothari09@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
Surface credential death to the event loop after dead-lettering the batch and posting the re-auth notice. Signed-off-by: Taksh <takshkothari09@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
Keeps the harness process up for logs/restart, but stops looking online in the sidebar once Claude/Codex credentials are dead (block#3831). Signed-off-by: Taksh <takshkothari09@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
Pin the headless Claude expiry message so a future precision tweak cannot quietly drop it from the non-retryable set. Signed-off-by: Taksh <takshkothari09@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
Regression pin for the block#3831 sidebar-online-after-credential-death path. Signed-off-by: Taksh <takshkothari09@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
The channel notice now matches presence behavior — re-login alone is not enough until the harness process is restarted. Signed-off-by: Taksh <takshkothari09@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
Operators running headless Claude agents hit this often; put the recovery steps next to the other harness docs. Signed-off-by: Taksh <takshkothari09@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
Contributor
Author
|
@tlongwell-block @wesbillman @wpfleger96 mind taking a look when you get a chance? |
|
Independent macOS managed-agent reproduction. The failure surfaced as: The current patch matches Reauthenticating the Claude subscription and restarting the managed runtime restored turns. Could this PR add a precise match and fail-first test for
This is the same user-facing failure as #3831, just a fourth message variant. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Authentication required(-32000) as a non-retryable auth error alongsideRe-authenticate/API Error: 401.offlineso the agent stops looking live in the sidebar.buzz-acpREADME.Fixes #3831
Test plan
cargo test -p buzz-acp --lib auth_errorclaude auth logout(or expire creds), @mention it — expect channel notice + sidebar goes offline, no further online heartbeatsMade with Cursor