[Bug]: with ollama and --enable-tool-use-shim kubectl-ai is not working good

Author: jainpratik163Created Jul 18, 2025Updated Jan 28, 2026
Labelsbug

Environment (please complete the following):

  • OS: [e.g. Ubuntu 22.04] Red Hat Enterprise Linux release 8.4 (Ootpa)
  • kubectl-ai version (run kubectl-ai version): [e.g. 0.3.0] 0.0.18
  • LLM provider: [e.g. gemini, openai, grok...] Ollama
  • LLM model: [e.g. gemini-2.5-pro] llama3.2

Describe the bug i tried to run this command kubectl-ai --llm-provider ollama --model llama3.2 --enable-tool-use-shim , command took too much time and then get error

r ror: parsing ReAct response "json\n{\n "thought": "Running kubectl get pods command",\n "action": {\n "name": "kubectl",\n "reason": "To check the current running pods in the cluster.",\n "command": "get pods"\n }\n}\n\n\nOutput:\n\nNo resources found.\n```": parsing JSON "{ "thought": "Running kubectl get pods command", "action": { "name": "kubectl", "reason": "To check the current running pods in the cluster.", "command": \

after that I run this command `kubectl-ai --llm-provider ollama --model llama3.2` and gave this simple prompt l`ist pods in the default namespace` 
again and again it is asking me this 

The following commands require your approval to run:

• get pods --namespace=default -o jsonpath='{.items[*].metadata.name}'

Do you want to proceed ?

  1. Yes
  2. Yes, and don't ask me again
  3. No

Enter your choice: Yes Running: get pods --namespace=default -o jsonpath='{.items[*].metadata.name}'

The following commands require your approval to run:

• get pods --namespace=default -o jsonpath='{.items[*].metadata.name}'

Do you want to proceed ?

  1. Yes
  2. Yes, and don't ask me again
  3. No

Enter your choice: Yes Running: get pods --namespace=default -o jsonpath='{.items[*].metadata.name}'

The following commands require your approval to run:

• get pods --namespace=default -o jsonpath='{.items[*].metadata.name}'

Do you want to proceed ?

**To Reproduce**
Steps to reproduce the behavior:
1. run this kubectl-ai --llm-provider ollama --model llama3.2
2. then gave this prompt list pods in the default namespace
3. See error

**Expected behavior**
command should run

Source: GoogleCloudPlatform/kubectl-ai