nocobase-ai-builder

Pass

Audited by Gen Agent Trust Hub on Aug 7, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill implements significant safety measures, including mandatory source backups or Git recovery points before upgrades, and explicit directives to avoid hardcoding secrets or credentials. It also requires reading project-specific instructions (AGENTS.md) before making changes. It identifies a surface for indirect prompt injection via ingestion of local source code and Swagger metadata from the NocoBase API, which is addressed through a structured design and verification workflow. Ingestion points include local files and API metadata; capability inventory includes file writes and CLI execution; sanitization is achieved through final build and browser verification steps.
  • [EXTERNAL_DOWNLOADS]: The skill manages source code via nb portal pull from configured repositories and visual foundations from shadcn.io/theme. These actions are central to the developer workflow and target the vendor's own infrastructure or established services.
  • [COMMAND_EXECUTION]: The skill invokes the NocoBase CLI (nb) for tasks such as pulling portal source, reading Swagger documentation, and querying collection metadata. These are necessary operations for designing and verifying a production-ready application.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 7, 2026, 07:11 PM
Security Audit — agent-trust-hub — nocobase-ai-builder