Skip to content

bug(a2a): LLMResponse repr leaking into system_commands agent artifact explanation field #4530

@mrveiss

Description

@mrveiss

Bug

The system_commands agent artifact explanation field contains a raw Python repr of LLMResponse instead of the actual response content:

"explanation": "LLMResponse(content='...',model='gemma2:2b', provider='ollama', ...)",

Root Cause

The same LLMResponse.get() / str-cast pattern fixed in #4501 for chat_agent.py also exists in the system_commands / EnhancedSystemCommandsAgent code path. The agent is likely doing str(llm_response) or passing the object directly instead of accessing .content.

Acceptance Criteria

  • explanation field contains the actual LLM response text
  • No LLMResponse( repr in any artifact field
  • System commands agent tests pass

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions