auth-tool-cloudbase

Pass

Audited by Gen Agent Trust Hub on Apr 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted external configuration data and using it within sensitive cloud management API payloads.
  • Ingestion points: Variables including environment IDs (env), WeChat credentials (AppID, AppSecret), and Google OAuth credentials (Client ID, Client Secret) are ingested from the agent context and used in SKILL.md.
  • Boundary markers: No explicit delimiters or boundary markers are defined to isolate user-supplied configuration values from the structure of the API request templates.
  • Capability inventory: The skill uses powerful service actions such as ModifyLoginConfig and ModifyProvider within the tcb and lowcode services, which can alter the security configuration of the environment.
  • Sanitization: The skill's instructions lack logic for sanitizing, validating, or escaping external inputs before they are interpolated into the JSON request bodies for management operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 3, 2026, 09:15 AM