What this category of data looks like
The data for this category primarily comes from code repositories, CI pipelines, Issue links, and contributor submission records. Daily associated documents include PR descriptions, code review comments, CI execution results, and merge records. Data update frequency is typically daily or weekly, driven by code commits, CI triggers, or merge operations. Fields include change scope, test pass rate, conflict markers, contributor identifiers, and more. There is no unified mandatory format, and it adjusts based on the configuration of the corresponding code repository.
Constraints for Tool Calling and Plugin Workflows
The data sources for this category are scattered and have inconsistent formats. Tool calling must adapt to parsing rules for multiple data types, including code repository APIs, CI log parsing, and Issue link redirection, to avoid call failures due to format differences. The data update frequency is daily or weekly, so tool trigger timing must match the rhythm of code commits, CI executions, or merge operations, and should not be too frequent to avoid occupying excessive system resources. Daily associated documents include code snippets and review comments, so the tool must support parsing of unstructured text and code blocks to ensure input content meets model processing requirements. Additionally, for use scenarios targeting project maintainers, tool calling should simplify configuration steps to lower the barrier to use.
Configuration Settings
The following configurations are based on FastGPT v4.17.0, and are adapted for the Pull Request review tool calling scenario.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
toolChoice | auto | Adapts to multi-tool calling scenarios for PR reviews, automatically matching tools for code parsing, CI result querying, and more |
functionCall | strict: true | Enforces parameter format validation for tool calls, ensuring input complies with preset review data parsing rules |
| Tool call trigger timing | 10–15 minutes after code commit | Matches the execution rhythm of CI pipelines, enabling timely retrieval of latest change and test results |
| Tool call timeout duration | 300 seconds | Adapts to the typical return duration of CI results, terminating calls upon timeout to avoid blocking review workflows |
| Number of tool call results to display | Top 3 entries | Focuses on core information required for PR reviews, filters redundant content, and improves review efficiency |
Common Misconfigurations
- An error stating "incorrect API Key used" is returned during calls. This occurs when global general-purpose keys are mistakenly used instead of application-specific keys when calling workflow tools.
- Tool call results include knowledge base search input and response content. This happens when non-essential knowledge base retrieval modules are not disabled in workflow configurations, causing unrelated content to be mixed into the tool call chain.
- No results are returned after tool call triggers. This is because the configured trigger timing occurs after the CI pipeline has completed, so the change data to be queried has not yet been generated.
How to Verify Configurations
- Submit a test PR to trigger the tool call chain, and check if the returned results only include core information required for reviews.
- Review tool call logs to confirm that the API key usage type matches the configured application permissions.
- Verify the trigger timing against the CI pipeline execution rhythm, confirming that complete change and test data is available when the call is made.
- Simulate the tool call parameter format to validate compliance with the
functionCallstrict validation rules.
The values provided here are common starting points, and should be evaluated 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.