The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the copaw command-line interface to manage scheduled tasks. It defines specific subcommands such as create, list, get, state, pause, resume, delete, and run to interact with a cron service.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to the way it handles user-provided content. Specifically, when creating a task of type agent, the content of the --text parameter is stored and later processed as a direct instruction to the agent.
Ingestion points: The --text parameter and the -f job_spec.json file input in the copaw cron create command within SKILL.md.
Boundary markers: Absent. The skill does not use delimiters or instructions to treat the user-supplied text as data rather than instructions.
Capability inventory: The skill allows creating (create), listing (list), and executing (run) tasks that can interact with other agents or channels.
Sanitization: Absent. There are no validation or escaping mechanisms described for the input parameters.

cron