ai-agents-openapi-agent-config-using-create-openapi-agent-results-in-valueerror-y-fb60d60e

THEN Modify the source file langchain_community/agent_toolkits/openapi/planner.py to pass allow_dangerous_requests=True when instantiating RequestsGetToolWithParsing and RequestsPostToolWithParsing. Alternatively, apply a monkey-patch or override the tool initialization. The parameter is not forwarded by create_openapi_agent, so direct modification is required until the library is fixed.

In langchain_community/agent_toolkits/openapi/planner.py, change:
tools: List[BaseTool] = [
    RequestsGetToolWithParsing(requests_wrapper=requests_wrapper, llm_chain=get_llm_chain),
    RequestsPostToolWithParsing(requests_wrapper=requests_wrapper, llm_chain=post_llm_chain),
]
to:
tools: List[BaseTool] = [
    RequestsGetToolWithParsing(requests_wrapper=requests_wrapper, llm_chain=get_llm_chain, allow_dangerous_requests=True),
    RequestsPostToolWithParsing(requests_wrapper=requests_wrapper, llm_chain=post_llm_chain, allow_dangerous_requests=True),
]