Security audit
Security checks for vulnerabilities and agentic risk
The skill is a workspace memory tool with clear safety boundaries, but it can be invoked automatically and persist session-derived learnings without explicit per-write consent.
Review this before installing if the workspace may contain sensitive operational details. The skill should only write sanitized summaries under the current workspace’s `.learnings/`, but users or administrators should consider disabling implicit invocation or requiring explicit save requests for privacy-sensitive agents.
The skill’s natural-language instructions and description are entirely in Chinese, and there is no indication that users may choose another language or that the skill is intentionally restricted to a Chinese-speaking or region-specific context. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.
The default_prompt is broad enough to trigger on common situations like recording corrections, failures, or reusable experience, which can cause the skill to activate in contexts the user did not explicitly request. Because this skill writes persistent data in the workspace, unintended activation can lead to over-collection of conversation content, incorrect memory formation, or silent persistence of sensitive operational details.
Enabling implicit invocation without narrow constraints allows the platform to auto-select this skill in a wide range of workflows. In this skill's context, that is more dangerous because it performs writes to .learnings/ and may persist user corrections, command failures, or session-derived content without sufficiently explicit consent, creating privacy, integrity, and prompt-injection persistence risks.
No suspicious patterns detected.
