letta-api-client

Pass

Audited by Gen Agent Trust Hub on Apr 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes implementation examples for client-side tool execution in examples/python/13_client_side_tools.py and examples/typescript/13_client_side_tools.ts. These scripts demonstrate how a local application can execute shell commands on behalf of an agent using subprocess.run(shell=True) and execSync. This is a documented architectural pattern of the Letta platform designed to grant agents access to local resources, and the examples emphasize the use of mandatory approval flags to ensure human oversight.
  • [EXTERNAL_DOWNLOADS]: The skill's documentation and examples reference the official vendor SDKs letta-client and @letta-ai/letta-client. It also includes examples of creating tools that fetch data from well-known third-party services such as the OpenWeatherMap API.
  • [SAFE]: All examples within the skill utilize secure secret management practices, employing environment variable placeholders rather than hardcoded credentials. Additionally, the documentation provides best-practice guidance for tool development, such as scoping library imports within function bodies to comply with sandboxed execution requirements.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 23, 2026, 06:38 AM