claude-code - 💡(How to fix) Fix [Bug] Auto-mode yields to user after Agent tool returns instead of continuing pipeline [3 comments, 3 participants]

[ { "error": "Error: Request was aborted.\n at makeRequest (/$bunfs/root/src/entrypoints/cli.js:50:3448)\n at processTicksAndRejections (native:7:39)", "timestamp": "2026-04-30T02:03:47.764Z" }, { "error": "MaxFileReadTokenExceededError: File content (27532 tokens) exceeds maximum allowed tokens (25000). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.\n at Km7 (/$bunfs/root/src/entrypoints/cli.js:4852:12789)\n at processTicksAndRejections (native:7:39)", "timestamp": "2026-04-30T02:50:21.614Z" }, { "error": "Error: EISDIR: illegal operation on a directory, read '/repos/<repo>/<module-dir>/test/<fixture-a>'\n at GAH (/$bunfs/root/src/entrypoints/cli.js:1708:144)\n at processTicksAndRejections (native:7:39)", "timestamp": "2026-04-30T02:50:44.245Z" }, { "error": "Error: EISDIR: illegal operation on a directory, read '/repos/<repo>/<module-dir>/test/<fixture-b>'\n at GAH (/$bunfs/root/src/entrypoints/cli.js:1708:144)\n at processTicksAndRejections (native:7:39)", "timestamp": "2026-04-30T02:50:44.432Z" } ]

TL;DR

Treat Agent tool returns as a strong "continue work" signal under auto mode to resolve the issue of the pipeline yielding to the user instead of continuing.

Guidance

• Review the pipeline implementation to ensure that the Agent tool returns are properly handled and that the pipeline is configured to continue execution after each step.
• Investigate the system-reminder mechanism for auto mode and consider making it more prominent, such as elevating it to a system-prompt level, to override the trained "tool result = end turn" instinct.
• Analyze the session JSONL bundle to identify patterns in the pipeline execution and determine if there are any specific conditions that trigger the pipeline to yield to the user.
• Consider implementing a bias toward continuation when the next step is enumerated in the skill body to improve the pipeline's autonomy.

Example

No code snippet is provided as the issue description does not include specific code references.

Notes

The issue may be related to the model's interpretation of long-running tool results as a natural turn end signal, which overrides the auto-mode and pipeline-state context. The provided error logs do not seem to be directly related to the issue, but may indicate other problems in the system.

Recommendation

Apply a workaround by treating Agent tool returns as a strong "continue work" signal under auto mode, as this is the most direct approach to resolving the issue. This change can help improve the pipeline's autonomy and reduce the need for user intervention.