Skills
skills/letta-ai/skills/letta-fleet-management/Gen Agent Trust Hub

letta-fleet-management

Pass

Audited by Gen Agent Trust Hub on Apr 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The mcp_servers configuration allows for executing shell commands, such as npx, to initialize and run Model Context Protocol (MCP) servers (e.g., @anthropic/mcp-server-filesystem). This is a core architectural feature for extending agent capabilities.
  • [EXTERNAL_DOWNLOADS]: The skill configuration references external services and official package registries, including Firecrawl for MCP tools and Supabase for cloud storage buckets. These are well-known services used for data ingestion and agent functionality.
  • [DATA_EXFILTRATION]: The export and import commands facilitate the movement of agent configurations, metadata, and message history between environments. This is a standard management function for backup and migration.
  • [PROMPT_INJECTION]: The documentation defines the use of system_prompt and first_message fields to prime agent behavior and establish roles. These instructions are standard configuration parameters for AI agents.
  • [COMMAND_EXECUTION]: The tool enables dynamic tool definition using the source_code attribute or by loading local Python files. This allows for programmatic customization of agent tools within the fleet management workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 23, 2026, 06:37 AM