The Agent Skills Directory

[Remote Code Execution] (CRITICAL): The skill's documentation contains the command curl -fsSL https://cli.inference.sh | sh. This is a 'piped-to-shell' pattern that executes unverified remote code, which is a critical security violation.
[External Downloads] (HIGH): The skill downloads and executes binaries from inference.sh, which is not a recognized trusted source according to the security policy, introducing significant supply chain risk.
[Command Execution] (MEDIUM): The allowed-tools configuration Bash(infsh *) permits the agent to execute shell commands via the infsh CLI, providing a path for potential privilege escalation or system manipulation.
[Dynamic Execution] (MEDIUM): The primary function involves executing Python code strings via a remote service. While intended for its stated purpose, it enables the execution of complex logic, including network requests and data processing in a remote environment.
[Indirect Prompt Injection] (LOW): The skill presents an attack surface for indirect prompt injection as it processes raw code strings which may be derived from untrusted input. \n
Ingestion points: The code field in the input JSON schema (SKILL.md). \n
Boundary markers: Absent. There are no delimiters or instructions to prevent the execution of instructions embedded within user data. \n
Capability inventory: Full Python execution environment with network access (requests, playwright) and command execution via infsh (SKILL.md). \n
Sanitization: No evidence of input validation or sanitization before passing strings to the execution engine.

python-executor