cloudbase-agent

Pass

Audited by Gen Agent Trust Hub on Nov 12, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation of AI agents that process untrusted user input via the AG-UI protocol. This inherent attack surface is managed through standard framework patterns such as LangGraph and LangChain.
  • Ingestion points: External data enters the agent context through the messages field in RunAgentInput, as documented in ts/agui-protocol.md.
  • Boundary markers: Developers are expected to define their own system prompts and boundary markers; the SDK itself provides the transport layer and does not enforce specific delimiters at the protocol level.
  • Capability inventory: The SDK supports LLM tool invocation, custom adapters, and state persistence in the CloudBase database.
  • Sanitization: Sanitization of message content is the responsibility of the implementing developer; the provided snippets demonstrate how to route and process messages without adding specific filtering.
  • [EXTERNAL_DOWNLOADS]: The documentation instructs users to install several official Tencent CloudBase packages (e.g., @cloudbase/agent-server) using the @latest version tag. This practice is documented as necessary for maintaining dependency alignment within the CloudBase environment and involves trusted vendor resources.
  • [COMMAND_EXECUTION]: Deployment instructions include the creation of a startup script (scf_bootstrap) and the use of chmod +x. These commands are standard for service deployment on the target platform and do not involve unauthorized privilege escalation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 12, 2026, 02:43 AM