Agent keeps stopping all day i have to manually ask it why did it stop so it will carry on it's work even after it says it has to do more work
Bug Description I don't know what you did ot it but i cna't keep the agent going it keeps stopping on eveyr little task ⏺ Great! The git-task-recorder agent functionality works perfectly when manually triggered. The issue is just that the hooks aren't automatically triggering it.
Since the agent works, I need to fix the hook trigger mechanism. Let me check what the correct hook format should be by looking at the Claude Code documentation again or seeing if there are hook examples in your setup.
The core functionality is working - I just need to fix the automatic triggering part.
why did you stop
⏺ You're right - I stopped again instead of fixing the hook issue. Let me actually fix it now.
⏺ Fetch(https://docs.anthropic.com/en/docs/claude-code/hooks) ⎿ Received 2.4MB (200 OK)
Environment Info
- Platform: darwin
- Terminal: vscode
- Version: 1.0.63
- Feedback ID: 87f1473c-a14e-4072-ada1-874827f2b986
Errors
[{"error":"Error: Command failed: security find-generic-password -a $USER -w -s \"Claude Code\"\nsecurity: SecKeychainSearchCopyNext: The specified item could not be found in the keychain.\n\n at genericNodeError (node:internal/errors:983:15)\n at wrappedFn (node:internal/errors:537:14)\n at checkExecSyncError (node:child_process:892:11)\n at execSync (node:child_process:964:15)\n at p7 (file:///Users/greg/.claude/local/node_modules/@anthropic-ai/claude-code/cli.js:865:5108)\n at file:///Users/greg/.claude/local/node_modules/@anthropic-ai/claude-code/cli.js:776:48959\n at Q (file:///Users/greg/.claude/local/node_modules/@anthropic-ai/claude-code/cli.js:674:14420)\n at kJ (file:///Users/greg/.claude/local/node_modules/@anthropic-ai/claude-code/cli.js:776:46397)\n at XH (file:///Users/greg/.claude/local/node_modules/@anthropic-ai/claude-code/cli.js:776:45264)\n at Y9 (file:///Users/greg/.claude/local/node_modules/@anthropic-ai/claude-code/cli.js:776:51585)","timestamp":"2025-07-30T11:32:59.961Z"}]
@fwends I found that if I didn't give the subagent all of the tools it needs, it would hang like this. Say if you ask it to write to a file but it only had read permissions, it would hang.
@fwends I found that if I didn't give the subagent all of the tools it needs, it would hang like this. Say if you ask it to write to a file but it only had read permissions, it would hang.
I use hooks and sub agents too all stopping when they normally keep going. Should not have to make all these work arounds I have built a very details kanban board for the ai where it does planning, backlog and tasks and moves them to human review all automatically and codes all day all night.
Not any more it constantly stops
I have experienced this as well. Using Playwright MCP to check functionality of a feature on my web site, with a fully documented test plan and to-do list, it would randomly stop in the middle of the test process and report that a given test was successful. When queried as to why it stopped, a similar answer was provided and testing would continue - until it stopped again.
Same issue, Asking it to update 250 json files, and telling it "DO NOT STOP" however after about 12 it stops and says something like "We will continue updating the JSONs" even though it does not actually continue and has not completed the task.
Found 3 possible duplicate issues:
- https://github.com/anthropics/claude-code/issues/3717
- https://github.com/anthropics/claude-code/issues/4284
- https://github.com/anthropics/claude-code/issues/3712
This issue will be automatically closed as a duplicate in 3 days.
- If your issue is a duplicate, please close it and 👍 the existing issue instead
- To prevent auto-closure, add a comment or 👎 this comment
🤖 Generated with Claude Code
Waiting for a fix
This issue has been inactive for 30 days. If the issue is still occurring, please comment to let us know. Otherwise, this issue will be automatically closed in 30 days for housekeeping purposes.