Describe the problem/error/question
Hi community, I’m trying to use my agent to recopile some info hitting two endpoints, with that reply I’m trying to send and email via gmail. The gmail node is not connected as a tool to the agent.
What is the error message (if any)?
Logs:
Problem in node ‘AI Agent‘
Model output doesn't fit required format
Structured Output Parser node logs:
Error in 1ms
Input
1 item
Error running node 'Structured Output Parser'
Output
1 item
Model output doesn't fit required format
To continue the execution when this happens, change the 'On Error' parameter in the root node's settings
Error details
Other info
n8n version
`1.122.4 (Self Hosted)`
Time
`12/8/2025, 12:38:34 PM`
Other info
n8n version
`1.122.4 (Self Hosted)`
Time
`12/8/2025, 12:38:34 PM`
Please share your workflow
My prompt is as follow:
“You are a cybersecurity agent.
You have access to two tools:
-
HybridAnalysisAPI_SearchHash
Use this tool to search for a hash and obtain the “report_id”.
Endpoint: /search/hash
The tool will return a JSON object containing “report_id” in its response.
-
HybridAnalysisAPI_ReportSummary
Use this tool to obtain the analysis summary.
You must call this tool using the report_id obtained previously.
Full endpoint required:
https://hybrid-analysis.com/api/v2/report/{report_id}/summary
IMPORTANT: Replace {report_id} in the URL with the actual value obtained from the first tool.
The hash to be analyzed is: {{ $(‘On form submission’).item.json.File_hash }}
Required workflow:
-
First, call HybridAnalysisAPI_SearchHash using the received hash. -
Read the “report_id” from the JSON returned by the first tool. -
Construct the URL for the second tool by replacing {report_id} with the obtained value. -
Call HybridAnalysisAPI_ReportSummary with that complete URL. -
Analyze the response from the second tool and return a summary with:
-
classification -
threat score -
indicators -
behavior
The final output must be plain text, clear, and concise. “
I selected the option ‘Require Specific Output Format‘ and added a “Structured Output Parser”
with this format:
Schema type: Define using json
{
“type”: “object”,
“properties”: {
“final_output”: {
“type”: “string”
}
},
“required”: \[“final_output”\]
}
The ouptut of that tools is:
Model output doesn’t fit required format
To continue the execution when this happens, change the ‘On Error’ parameter in the root node’s settings
And my Agent is not giving any output, all the reasoning is on the input but it do not write the output, so in my next node(gmail) I cannot use the output of the agent
Information on your n8n setup
- n8n version: Version 1.122.4
- Database (default: SQLite): Default
- n8n EXECUTIONS_PROCESS setting (default: own, main):
- Running n8n via (Docker, npm, n8n cloud, desktop app): Docker
-
Operating system: ubuntu 24