The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The documentation includes an installation command that fetches a shell script from https://cli.inference.sh and pipes it to sh. This is the vendor's official distribution method for their command-line interface.
[COMMAND_EXECUTION]: The skill uses the infsh CLI tool to run applications on the vendor's platform. The execution environment is restricted by the platform to only allow infsh commands, which limits the potential attack surface.
[INDIRECT_PROMPT_INJECTION]: The skill accepts arbitrary Python code for execution, creating a potential surface for indirect injection if processing untrusted data.
Ingestion points: The code field in the input JSON (SKILL.md).
Boundary markers: Absent.
Capability inventory: Remote code execution via infsh app run (SKILL.md).
Sanitization: Absent; the skill relies on the remote environment's sandbox for isolation.

python-executor