The Agent Skills Directory

[COMMAND_EXECUTION]: The skill allows execution of arbitrary shell commands through the execute_command tool. This includes environment inspection, system utility usage, and software installation.
[DYNAMIC_EXECUTION]: The skill is designed for the runtime execution of Python, JavaScript, and TypeScript code, enabling prototyping, debugging, and data processing within a persistent workspace.
[EXTERNAL_DOWNLOADS]: The environment supports fetching external resources via tools like curl and Python libraries such as requests. It also allows the installation of third-party packages from registries like PyPI.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it is designed to ingest and process data from external sources (web scraping, API calls, and user-uploaded files) that may contain hidden instructions.
Ingestion points: User-provided file attachments located in /mnt/workspace/inputs and external data fetched via network operations (requests, httpx).
Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore potentially malicious content embedded in processed data.
Capability inventory: The skill provides comprehensive system access including arbitrary code execution, shell command execution, and file system modification.
Sanitization: There are no explicit mechanisms or instructions for sanitizing, escaping, or validating external content before it is processed by the agent.

code-interpreter