Security audit
Security checks for vulnerabilities and agentic risk
This is a design-system guidance skill with broad edit and command permissions, but the inspected artifact shows no hidden payload, persistence, credential use, or data exfiltration.
Use this skill when you want an agent to inspect and modify a frontend project for design-system integration. Review proposed file edits and shell commands before applying them, especially npm installs, builds, tests, or network troubleshooting commands.
The skill declares exec alongside read and write even though its stated purpose is design-system guidance and token/configuration advice. In an agent environment, unnecessary shell execution materially expands the attack surface because natural-language prompts or future skill logic could trigger command execution unrelated to the design task.
The skill exposes execution and file-writing capabilities but does not provide a clear warning that it may modify project files or run shell commands. In practice, this can lead users or calling agents to invoke the skill without informed consent, enabling unintended file changes or command execution in the local environment.
The capability coverage text advertises a very broad trigger scope, including generic coding, debugging, testing, deployment, and automation scenarios well beyond a minimalist design-system skill. Overbroad matching increases the chance the skill is auto-invoked in unrelated development conversations, which becomes risky because the skill also has write/exec capabilities.
displayName、summary 和 description 均固定为中文表述，全文说明也默认以中文呈现，没有声明可根据用户偏好切换语言。根据语言/地区策略，若技能默认强制单一语言而不给用户选择，可能构成自然语言政策问题。
No suspicious patterns detected.
