Google ADK Python LoadSkillResourceTool Infinite Retry Loop Consumes Call Budget

What this error means

RESOURCE_NOT_FOUND — LoadSkillResourceTool retries indefinitely until max_llm_calls (default 500) exhausted is a Google ADK Python failure pattern reported for developers trying to fix google adk python loadskillresourcetool retry loop consuming entire call budget on resource_not_found. Based on the imported evidence, treat this as a tool-specific troubleshooting page rather than a generic API error.

Why this happens

GitHub issue #5652 reports LoadSkillResourceTool returning RESOURCE_NOT_FOUND as structured soft-error, causing LLM to retry same path until RunConfig.max_llm_calls (default 500) is exhausted. Single hallucinated path silently consumes entire per-invocation call budget.

Common causes

• When an LLM passes a hallucinated path to LoadSkillResourceTool in Google ADK Python, the tool returns RESOURCE_NOT_FOUND as a soft-error string. The LLM treats this as recoverable and retries indefinitely until max_llm_calls (default 500) is exhausted — silently consuming the entire call budget on a single failing tool.
• GitHub issue #5652 reports LoadSkillResourceTool returning RESOURCE_NOT_FOUND as structured soft-error, causing LLM to retry same path until RunConfig.max_llm_calls (default 500) is exhausted. Single hallucinated path silently consumes entire per-invocation call budget.

Quick fixes

1. Confirm the exact error signature matches RESOURCE_NOT_FOUND — LoadSkillResourceTool retries indefinitely until max_llm_calls (default 500) exhausted.
2. Check the Google ADK Python account, local tool state, and provider configuration involved in the failing workflow.
3. Compare the failing environment with a known working setup, then change one configuration value at a time.

Platform/tool-specific checks

• Verify the command, editor, extension, or API client that produced the error.
• Compare local settings with CI, deployment, or editor-level settings when the error appears in only one environment.
• Avoid deleting credentials, local model data, or project settings until the failing scope is clear.

Step-by-step troubleshooting

1. Capture the exact error message and the command, editor action, or request that triggered it.
2. Check whether the failure is account/auth, quota/rate, model/provider, local runtime, or deployment configuration.
3. Review the source evidence below and compare it with your environment.
4. Apply one change at a time and rerun the smallest failing action.
5. Keep the working fix documented for the team or deployment environment.

How to prevent it

• Keep provider/tool configuration documented.
• Record non-secret diagnostics such as tool version, provider name, model name, and command path.
• Add a lightweight check before CI or production workflows depend on the tool.

Sources checked

• https://github.com/google/adk-python/issues/5652

Evidence note: GitHub issue #5652 reports LoadSkillResourceTool returning RESOURCE_NOT_FOUND as structured soft-error, causing LLM to retry same path until RunConfig.max_llm_calls (default 500) is exhausted. Single hallucinated path silently consumes entire per-invocation call budget.

Related errors

• Google ADK Python tool retry loop bug
• ADK max_llm_calls budget exhaustion
• Google ADK skill resource path not found

FAQ

What should I check first?

Start with the exact RESOURCE_NOT_FOUND — LoadSkillResourceTool retries indefinitely until max_llm_calls (default 500) exhausted text and the smallest action that reproduces it.

Can I ignore this error?

No. Treat it as a failed Google ADK Python workflow until the root cause is understood.

Is this guaranteed to have one fix?

No. The imported evidence supports the troubleshooting path above, but tool behavior can vary by account, plan, version, provider, and local configuration.

How do I know the fix worked?

Rerun the same command, editor action, or request. The fix is working when that action completes without RESOURCE_NOT_FOUND — LoadSkillResourceTool retries indefinitely until max_llm_calls (default 500) exhausted.