cloudbase-platform
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it retrieves and processes data from external CloudBase environments and user-controlled configurations.\n
- Ingestion points: Data is ingested via management and query tools (e.g.,
queryGateway,queryHosting,queryStorage,manageFunctions) that retrieve state from the remote cloud environment, as documented inSKILL.md.\n - Boundary markers: The mandatory
Change Safety Protocolinreferences/protocols/change-safety-protocol.mdenforces a manual user confirmation step before the agent executes code or configuration changes based on ingested data.\n - Capability inventory: The skill possesses broad capabilities to modify cloud infrastructure, including environment management, function deployment, and database permission updates using the tools listed in the activation contract and management sections of
SKILL.md.\n - Sanitization: Robust sanitization rules are defined in the
Sensitive Runtime Data Protectionprotocol (references/protocols/sensitive-runtime-data-protection.md), which explicitly prohibits the reflection of sensitive platform headers (likex-cloudbase-context) and credentials to external clients.
Audit Metadata