The Agent Skills Directory

[PROMPT_INJECTION]: The skill exposes a code execution surface through the code parameter. While this is the primary purpose of the skill, it represents a potential vector for indirect prompt injection if the agent interpolates untrusted data into the code block.
Ingestion points: The code field in the input schema defined in SKILL.md.
Boundary markers: None present; there are no instructions to use delimiters or 'ignore' instructions within the interpolated code.
Capability inventory: Extensive Python runtime capabilities including network access (requests), browser automation (playwright), and file output management.
Sanitization: No sanitization or validation logic is specified within the skill instructions; execution safety relies on the remote platform's sandbox.
[EXTERNAL_DOWNLOADS]: The skill references external resources necessary for its operation.
Fetches setup documentation from the platform's GitHub repository: https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md.
Recommends installation of the platform-specific CLI via npx skills add belt-sh/cli.
[COMMAND_EXECUTION]: The skill invokes the belt CLI tool to perform code execution tasks. This behavior is consistent with the skill's documented purpose and uses the platform's intended command-line interface.

python-executor