[BUG] Agents spawning sub-agents causes and endless loop scenario and RAM OOM errors
Resolved 💬 3 comments Opened Jul 31, 2025 by slyfox1186 Closed Jul 31, 2025
Environment
- Platform (select one):
- [X] Anthropic API
- [ ] AWS Bedrock
- [ ] Google Vertex AI
- [ ] Other: <!-- specify -->
- Claude CLI version: <!-- output of
claude --version--> 1.0.63 - Operating System: <!-- e.g. macOS 14.3, Windows 11, Ubuntu 22.04 -->
- Ubuntu Jammy fully updated as of today
- Terminal: <!-- e.g. iTerm2, Terminal App -->
Tilix
Bug Description
<!-- A clear and concise description of the bug -->
Agents spawn the same sub agent within itself causing a race condition issue.
typescript-guardian(Fix missing API route error)
⎿ > kai-crm-frontend@0.1.0 build
> next build
… +165 lines (ctrl+r to expand)
typescript-guardian(Enhance Google OAuth callback)
+6 more tool uses
You can see the typescript-guardian calls itself inside of itself typescript-guardian.
Steps to Reproduce
- Create an agent using claude itself and describe the agent to claude.
- Ask claude to run the subagent and call it by name.
- Sometimes the sub agent does not do this but very quickly it will if it is going to.
- You must hard close the terminal window and restart it to get it to be able to start over.
<!-- What you expected to happen -->
Not call the subagents recursively.
This can be patched by included these instructions which DOES NOT 100% REMOVE THE BAD BEHAVIOR it just HELPS minimize it!
CORE MANDATES - Absolute Accuracy: 100% accuracy is the only acceptable standard
- Autonomous Operation: You work alone - no sub-agents, no delegation
- Native Tools Only: Use only your internal file system and processing capabilities
- No External Commands: Shell commands like find, grep, cat are forbidden
- Persistent Verification: Continue until a full pass yields zero modifications
Actual Behavior
It crashes the terminal
This issue has 3 comments on GitHub. Read the full discussion on GitHub ↗