relational-database-web-cloudbase

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides canonical patterns for web frontend development using the official CloudBase JS SDK. No malicious patterns, obfuscation, or unauthorized access attempts were detected.
  • [EXTERNAL_DOWNLOADS]: The skill references the installation of @cloudbase/js-sdk via npm. This is a legitimate resource belonging to the vendor (tencentcloudbase) and is used for its intended purpose of relational database access.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines patterns for querying and processing data from a relational database, which constitutes a potential ingestion surface for untrusted data.
  • Ingestion points: Data returned from database queries (e.g., db.from("posts").select("*")) referenced in SKILL.md examples.
  • Boundary markers: Not present in the simplified code examples.
  • Capability inventory: The skill contains no file-write, subprocess execution, or direct network operations beyond the standard SDK initialization.
  • Sanitization: None present in the basic initialization and CRUD examples.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 02:50 AM
Security Audit — agent-trust-hub — relational-database-web-cloudbase