Security audit
Security checks for vulnerabilities and agentic risk
The skill is a coherent XXL-JOB debugging guide, but it gives agents authority to kill and delete scheduler jobs without requiring explicit final confirmation.
Install only for users who intentionally manage XXL-JOB environments. Use least-privileged tokens and prefer development or isolated schedulers; before any kill, delete, start, stop, update, or production trigger, require the agent to show the exact jobId, appName, executor, and expected impact and get explicit confirmation.
The skill documents executor-side kill operations and says to use them on explicit request or isolated debugging, but it does not mandate a user-facing warning or confirmation before terminating a running task. In scheduler/debug workflows, killing a live job can interrupt business processing, leave partial state, and cause data inconsistency if done on the wrong executor or job.
The skill includes deletion of XXL-JOB tasks and even recommends deleting temporary debug tasks, but it does not require an explicit confirmation or prominent warning immediately before destructive removal. In an operational scheduler context, this can cause accidental deletion of real jobs, loss of configuration, and service disruption if an agent interprets context incorrectly or the user request is ambiguous.
The skill uses Chinese natural-language examples and later repeats Chinese-only trigger phrasings for operation mapping, but does not state that other languages are supported or give the user a language/locale choice. This can be a language policy issue because the skill appears to assume a specific language without opt-in or justification.
No suspicious patterns detected.
