agent-tools

Pass

Audited by Gen Agent Trust Hub on Apr 24, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill documents and instructs the user/agent to install the official infsh CLI using a piped-to-shell command: curl -fsSL https://cli.inference.sh | sh. This is the standard, official installation method for the inference.sh platform.
  • [COMMAND_EXECUTION]: The skill uses the Bash(infsh *) tool to execute commands on the host system. This is limited to the infsh CLI tool provided by the vendor.
  • [DATA_EXFILTRATION]: The CLI tool includes a feature to upload local files (e.g., images, audio) to the cloud for processing by AI models. This is a documented core capability of the service and occurs only when the agent or user explicitly provides a local file path to an app execution command.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 24, 2026, 08:29 AM
Security Audit — agent-trust-hub — agent-tools