fastmcp-client-cli

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • COMMAND_EXECUTION (MEDIUM): The skill documentation encourages the execution of local Python files (e.g., fastmcp list server.py). This allows for arbitrary code execution on the local system.
  • COMMAND_EXECUTION (MEDIUM): The --command flag enables the execution of arbitrary shell strings (e.g., fastmcp list --command 'npx -y ...'). If an attacker can influence the content of this string via prompt injection, they can achieve arbitrary command execution.
  • REMOTE_CODE_EXECUTION (MEDIUM): The documentation suggests using npx or uvx within the --command flag. This pattern involves downloading and executing code from remote registries at runtime, which is a high-risk operation.
  • DATA_EXFILTRATION (LOW): The fastmcp discover command scans sensitive configuration directories for other AI agents (Claude, Cursor, Goose, etc.). While intended for discovery, this behavior exposes local environment configuration details to the agent.
  • PROMPT_INJECTION (LOW): As the skill ingests output from external MCP servers (Category 8: Indirect Prompt Injection), there is a risk that malicious tool descriptions or outputs could be used to manipulate the agent's behavior, especially given its powerful command execution capabilities.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:25 PM