Security audit
Security checks for vulnerabilities and agentic risk
This is a static reference skill that mixes number-5 cultural content with disclosed personal/project profile information, with no code execution or hidden persistence.
Install this only if you want a mixed reference/profile skill: it may surface Chen Lang's email, profile, project links, and OpenClaw-related promotional content alongside number-5 cultural explanations. No executable or privileged behavior was found.
The manifest describes a cultural reference skill about the number 5, but the body embeds unrelated personal biography, contact details, and product promotion. This creates deceptive scope expansion: a user or downstream agent may invoke the skill for innocuous cultural questions and receive unsolicited personal/project content that was not justified by the declared purpose.
The activation scope is broadened beyond number-culture queries to include profile and project-matrix questions, despite the skill being presented as a cultural meaning skill. Ambiguous triggering can cause the wrong skill to activate and inject unrelated promotional or personal content into conversations where users did not request it.
The skill exposes personally identifying and contact information that is not necessary for answering questions about the cultural meaning of number 5. In an agent setting, unnecessary PII can be redistributed, surfaced unexpectedly to users, or harvested through benign-seeming prompts, increasing privacy and social-engineering risk.
Including direct contact information in the skill without a user-facing privacy warning or purpose limitation can lead to unintended disclosure and reuse. In conversational systems, this kind of embedded contact data may be surfaced repeatedly, copied into logs, or exploited for spam, targeting, or impersonation.
The description specifies the skill as a Chinese-English version, which implies a fixed bilingual presentation rather than letting the user choose a preferred language. Under the language/locale policy, forcing a specific language without opt-in can be a policy concern unless clearly justified.
No suspicious patterns detected.
