python-executor

Fail

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to download a setup script from the vendor's official domain at cli.inference.sh during the initial setup process.
  • [REMOTE_CODE_EXECUTION]: The installation procedure involves piping a remote script directly to a shell (curl | sh) to install the vendor's command-line interface. This is a common pattern for the installation of developer tools from trusted service providers.
  • [COMMAND_EXECUTION]: The skill utilizes the infsh CLI tool to facilitate the execution of Python scripts on the vendor's remote infrastructure.
  • [PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it accepts and executes arbitrary Python code provided as input.
  • Ingestion points: The code field within the input schema defined in SKILL.md.
  • Boundary markers: No specific delimiters or warnings to ignore instructions within the code are implemented in the skill's instructions.
  • Capability inventory: The skill allows for remote code execution with access to numerous libraries for networking and data processing via the infsh app run command.
  • Sanitization: No client-side sanitization or validation of the input code is performed by the skill before transmission to the remote service.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 8, 2026, 02:51 AM