Security audit
Security checks for vulnerabilities and agentic risk
This skill coherently creates OpenClaw scheduled tasks, but it can create persistent jobs and force immediate execution without requiring explicit user confirmation.
Review each cron command before allowing it to run. Confirm the schedule, agent, delivery target, webhook URL, and whether testing will immediately execute the task; use one-shot deletion for reminders and avoid script-backed recurring jobs unless you intentionally want ongoing automation.
The invocation description includes generic phrases like "schedule a task," "create a reminder," and "run something at a specific time," which are common everyday requests and may cause the skill to match too broadly. The description does not provide exclusion conditions or negative examples to clarify when this skill should not activate.
The skill instructs the agent to create persistent scheduled tasks and to immediately test them with openclaw cron run <jobId> without requiring an explicit user confirmation or warning about side effects. In this context, that is dangerous because cron jobs can create ongoing autonomous behavior, trigger notifications/webhooks, run scripts repeatedly, and execute immediately, turning an ambiguous or casually phrased request into durable actions with potentially significant operational impact.
No suspicious patterns detected.
