What data for this category looks like
Data for this category comes from the regulatory and rule repository, upper-level law database, departmental policy document repository, and administrative regulatory document filing system. Update frequency is irregular. Core daily processed documents include five types: draft regulatory documents, drafting explanations, legality review opinions, summary of solicited comments, and filing reports.
Core fields of documents revolve around document making authority, article content, and relevant source basis. Draft documents include full article text and drafting unit information. Review opinion documents include review conclusions and basis entries. Field boundaries for each material are clear, facilitating classified processing and review.
What constraints do these characteristics impose on the forms and interactions link
The data characteristics of this category impose multiple constraints on the forms and interactions link. Multi-source heterogeneous official data sources require forms to support multi-module associated input, allowing users to upload different types of review materials separately. Irregularly updated data sources require the interaction link to support an entry for real-time fetching of the latest basis materials, avoiding use of outdated regulatory articles. The structure of multiple independent documents requires the form to split input areas by material type, reducing user confusion. Strict compliance requirements require the interaction link to include permission control and required field validation, preventing unauthorized operations or missing core information and ensuring compliance of the review process.
Configuration settings
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
formInput | Split modules by review workflow: draft documents, basis materials, draft review opinions | This category of review materials includes multiple independent document types; splitting improves interaction clarity |
userSelect | Restricted to legal review-related positions within the current organization | Internal office work requires following permission control to prevent unauthorized personnel from operating review processes |
| Maximum number of files submitted per single submission | No more than 6 | Adapts to the batch processing rhythm of the review workflow, reduces risk of interaction lag |
| Maximum character limit per single input | 800–1200 characters | Adapts to the length of review text for single articles, avoids loading delays caused by overly long input content |
| Real-time basis fetching | Enabled | Data sources are irregularly updated official regulatory repositories; real-time fetching ensures timeliness of basis materials |
| Required field validation | Enable draft article and basis source fields | Missing core fields will prevent completion of the review, which meets compliance requirements |
Three common misconfigurations
- Custom review basis plugin does not display input parameters. Cause: No trigger binding for the corresponding plugin is configured in
formInput, causing the workflow task to fail to recognize the parameter entry. - Incorrect passing of array elements when cyclically calling AI to review single articles. Cause: The array output by the loop body is not bound to the dynamic parameters of
userSelect, causing the input end to fail to recognize the entries to be reviewed. - Preset review guidance prompts do not take effect in AI conversations. Cause: The guidance prompt synchronization switch is not enabled in the form configuration, causing the AI conversation to fail to load the preset review rule guidance.
How to confirm the configuration is successful
- Submit a test draft document, check if the form automatically triggers associated fetching of basis materials.
- Select a test account from a non-authorized position, check if access to this review form is blocked.
- Upload more than 6 test files, check if a submission limit prompt pops up.
- After enabling the guidance prompt configuration, launch an AI conversation, check if the preset review rules are loaded.
Note: The values provided are common starting points and should be evaluated against your 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 2026-09-28.