landbot-automation

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions direct the agent to execute automation tasks using RUBE_MULTI_EXECUTE_TOOL and RUBE_REMOTE_WORKBENCH. These tools are dynamically identified through runtime discovery.- [DATA_EXFILTRATION]: The skill requires network communication with rube.app and composio.dev to manage toolkit connections and retrieve current tool schemas. This behavior is documented and consistent with the intended automation functionality.- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes external data from Landbot and Composio tool outputs without explicit sanitization or boundary markers.\n
  • Ingestion points: Untrusted data enters the context through tool execution results and dynamic schema searches (SKILL.md).\n
  • Boundary markers: None identified in the instructions to separate external data from agent instructions.\n
  • Capability inventory: The skill has access to remote execution and workbench tools (RUBE_MULTI_EXECUTE_TOOL, RUBE_REMOTE_WORKBENCH).\n
  • Sanitization: No evidence of input validation or output sanitization is present in the provided instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 04:34 AM