The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides documentation for installing the jo utility from official package repositories and the project's GitHub repository (github.com/jpmens/jo). The instructions target a well-known open-source tool and follow standard installation procedures.
[COMMAND_EXECUTION]: The instructions involve executing the jo command-line utility, along with curl and kubectl, for constructing and sending JSON payloads. These examples are for legitimate developer workflows.
[INDIRECT_PROMPT_INJECTION]: The skill identifies a pattern for processing shell variables into JSON payloads, representing a standard surface for data processing.
Ingestion points: User-defined shell variables such as $NAME, $EMAIL, and $TOKEN are interpolated into jo commands within the examples in SKILL.md.
Boundary markers: No specific boundary markers are used for input variables, but jo itself serves as a structural boundary for the resulting JSON.
Capability inventory: Shell execution of jo, curl, and kubectl as documented in SKILL.md.
Sanitization: The primary purpose of the skill is to use jo as a sanitization layer, ensuring that user-provided variables are correctly quoted and escaped to prevent structural JSON injection.

jo