Data for This Category
Data for this category originates from server and middleware configuration files, security device policy logs, and dedicated baseline verification tools. Data update frequency is monthly or quarterly. Core documents include baseline checklists, configuration verification records, rectification suggestions, and exception approval forms. Baseline checklists contain configuration item names, current configuration values, and compliance baseline values. Configuration verification records record verification timestamps, lists of abnormal items, and executor information. Exception approval forms link the application reason and approval status for corresponding abnormal items.
Constraints for Tool Calling and Plugins
Since data sources are scattered across server and middleware configuration files, security device policies, and dedicated baseline verification tools, tool calling must support parsing and access to multi-format configuration files. The monthly or quarterly update frequency means high-frequency polling is unnecessary. Fixed-cycle task triggering can be configured instead. The structure of multiple associated documents requires plugins to support cross-document field association. For example, synchronizing the approval status from an exception approval form to the corresponding abnormal item in a baseline checklist. The high time cost of batch configuration item verification requires tool calling to support batch pulling of abnormal items. It must also include compliance status verification logic to prevent weak configurations from remaining in place for long periods. Cross-team knowledge base association scenarios require plugins to support cross-team permission verification. This ensures secure access to compliance baseline data.
How to Configure
This configuration applies to FastGPT v4.17.0
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
toolChoice | auto | Supports automatic calling of various baseline verification tools, no manual tool specification required |
functionCall | Set based on actual testing | Supports parameter passing for custom baseline check scripts |
| Global variable passing | Enable and bind workflow token variables | Meets identity credential transfer requirements across tools |
| Local MCP calling | Enable local service listening port | Supports direct calling of locally deployed baseline verification tools |
| Knowledge base associated queries | Bind compliance baseline knowledge base | Enables automatic matching of abnormal items and generation of rectification suggestions |
| Single abnormal item pull | Top 5 entries | Adapts to the batch verification rhythm of baseline checks, reduces single call load |
| Call timeout | 600 seconds | Accommodates reasonable time for pulling data from multiple sources |
Three Common Mistakes
- When calling an MCP tool in a workflow, global variables cannot be retrieved. The symptom is that the token field received by the tool is empty. The cause is that the global variable binding configuration is not enabled, and the workflow token is not mapped to the tool call parameters.
- Calling a local MCP tool fails. The symptom is a connection timeout error. The cause is that the local service listening port is not enabled, or a firewall blocks the tool communication port.
- Format corruption occurs when processing Word documents with revision marks. The symptom is lost revision content or messy typesetting. The cause is that a dedicated document parsing plugin is not used, and a general text parsing tool is employed to process formatted Word files.
How to Confirm Configuration Is Complete
- Trigger a workflow call, check the tool call log, and confirm that the global variable token has been correctly passed to the MCP tool parameters.
- Check the service status of the local MCP tool, confirm that the listening port is enabled, and that a call request can be initiated via the local address.
- Upload a Word document with revision marks, call the document processing plugin, and confirm that the original format and revision content are retained after parsing.
- Test associating a knowledge base from a different team, submit a baseline-related query, and confirm that the returned results include compliance baseline data.
The values provided are common starting points and should be measured against the reader's own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-10-19.