This document applies to FastGPT v4.17.0.
What data for this category looks like
Data for this category primarily comes from crowdsourced platforms, annotation quality data, training exam records, and financial systems. Daily maintained documents include personnel access rules, annotation training materials, performance ranking sheets, and settlement lists. Data update cadence follows daily and monthly cycles. Daily data includes daily annotation completion volume and quality inspection pass rate. Monthly data covers personnel performance summaries and settlement details. Data fields include unique personnel identifiers, annotation task types, quality inspection results, settlement amounts, and more. Association of personnel and task information is required to complete management.
What constraints do these characteristics impose on permissions and data isolation
Data sources span multiple systems, so access permissions for different systems must be isolated to prevent annotators from accessing sensitive data such as financial settlement information. Update frequencies are split into daily and monthly, so permission validity periods must be configured according to update cycles. Daily data requires real-time or short-cycle synchronization of permission scopes, while monthly data can have permission rules updated periodically. Documents contain sensitive content such as access rules and performance sheets, so access levels must be differentiated by role. For example, quality inspection supervisors can view performance sheets, while regular annotators can only view their own training materials. Personnel data is linked to annotation tasks, so permissions must be bound to task scopes to prevent cross-project data leaks.
How to set configurations
| Configuration Item | Recommended Approach | Rationale |
|---|---|---|
teamId | Assign unique identifiers per annotation project | Distinguish personnel and data boundaries between different annotation projects, preventing cross-project data access |
tmbId | Bind permission groups to personnel roles | Match the different responsibility scopes of annotators, quality inspection supervisors, and project leads |
| Permission scope configuration | Divide access levels by document type | Correspond to the varying sensitivity levels of personnel access rules, performance sheets, and settlement lists |
| Data synchronization cycle | Synchronize daily data every 600 seconds, monthly data every 86400 seconds | Adapt to the daily and monthly update rhythms of data, ensuring permission timeliness |
| Member account verification | Only allow adding accounts registered within the platform | Resolve issues with invalid usernames, ensuring member permissions are legitimate |
| Access audit retention | Retain for 30 days | Meet internal management and quality traceability requirements |
Three common configuration errors
- Phenomenon: After entering an external account in the team member addition interface, the interface prompts that the username is invalid. Cause: No verification of platform-registered accounts was performed, and non-platform members were attempted to be added directly.
- Phenomenon: Personnel from different annotation projects can view each other's performance and settlement data. Cause: Data access scopes were not bound by
teamId, and permission configurations did not distinguish project boundaries. - Phenomenon: After configuring permissions, internal ticket Q&A still returns unauthorized document content. Cause: Permission rules were not divided into access levels by
tmbId, and the configuration did not take effect.
How to confirm the configuration is complete
- Perform the team member addition operation, enter an account registered within the platform, and confirm that no invalid username prompt appears.
- Switch accounts with different roles, verify that only documents within the corresponding permission scope can be accessed. For example, annotators cannot view settlement lists.
- View the permission configuration log, confirm that the binding rules for
teamIdandtmbIdhave been correctly synchronized. - Initiate a cross-project ticket query, confirm that relevant data from other annotation projects cannot be obtained.
The values provided here 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.