python-executor
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references installation guidelines and CLI tools hosted on the official GitHub repository for the inference.sh platform.
- [COMMAND_EXECUTION]: The skill utilizes the
beltCLI tool to communicate with the sandboxed execution environment. This is controlled via theallowed-tools: Bash(belt *)configuration. - [DYNAMIC_EXECUTION]: The primary function of the skill is to execute arbitrary Python code provided in the input schema. This code is run in an isolated cloud environment.
- [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it ingests and executes code that may be generated based on untrusted external data.
- Ingestion points: The
codeparameter in the input schema (SKILL.md). - Boundary markers: None explicitly provided within the input code block.
- Capability inventory: Execution of Python code with network access (via
requests,selenium, etc.) and file system access within the sandbox (SKILL.md). - Sanitization: Relies on the platform's sandboxing mechanisms to prevent local system compromise.
Audit Metadata