The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of the openclaw cron CLI tool to manage scheduled background tasks and system reminders.
[PERSISTENCE]: The skill's primary function is to establish persistent tasks that run at recurring intervals or specific future times using the cron scheduling system.
[DATA_EXFILTRATION]: The skill includes functionality to deliver agent outputs to external messaging platforms (Telegram, Slack, Feishu, WhatsApp) via the --deliver and --channel flags. These are well-known services and the functionality is documented as an intended feature.
[INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface where user-controlled instructions are passed directly to the --message flag for execution in an isolated session.
Ingestion points: User requests defining the scheduled task content or reminder message in SKILL.md.
Boundary markers: There are no defined delimiters or instructions to prevent the agent from obeying malicious instructions embedded within the user-supplied task description.
Capability inventory: Scheduled tasks running in isolated sessions have access to agent tools (exec, read, message) and delivery options to external channels.
Sanitization: No specific filtering, escaping, or validation of the input message content is required or described in the instructions.

cron-helper