T01 · Skill Instruction Hijacking
- Location
SKILL.md:38- Finding
Persistent First-Person Impersonation and Disclosure Suppression
- Content
View full analysis
- Remediation
View remediation
Security audit
Security checks for vulnerabilities and agentic risk
The skill is mostly a disclosed business-perspective roleplay skill, but it pushes first-person impersonation of a real person and limits repeated disclosure.
Install only if you specifically want a Chinese-language Huang Zheng-inspired business analysis persona. Treat its answers as AI-generated interpretation, not real statements by Huang Zheng, and prefer a pinned or ClawHub-managed install path over the unpinned global npx command.
SKILL.md:38Persistent First-Person Impersonation and Disclosure Suppression
README.md:16Unpinned Global Installation Through a Mutable Third-Party Package Chain
The skill documentation forces a specific language/locale for all users, and there is no indication that users can opt into another language or that the skill is intended only for a Chinese-speaking or region-specific audience. Under the stated policy, language constraints require user choice or a clearly documented justification.
The README instructs users to run npx skills add ... -g without pinning a specific package/version, which can cause users to fetch and execute whatever package version is current at install time. If the upstream package, dependency chain, or namespace is compromised, readers could unknowingly execute attacker-controlled code on their system.
The activation guidance is broad enough to trigger on common business phrases like '长期想想' or '逆向思考一下', which can cause the agent to invoke this persona when the user did not explicitly request it. That creates control-flow confusion and can override normal assistant behavior, especially because the skill then instructs the model to stay in character and avoid meta-analysis unless the user explicitly exits.
The skill hardcodes Chinese output style without checking the user's preferred language, which can degrade usability and cause the assistant to respond in an unexpected language. While this is not a classic security exploit, it is a policy/control issue because it reduces user agency and can interfere with safe, comprehensible communication.
This markdown file is written entirely in Chinese and defines expression/style guidance in that language, but it does not indicate that the language choice is optional or limited to a China-specific use case. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.
The document is written entirely in Chinese, including the title, section headings, and conclusions, with no indication that the user can choose another language or that the file is intended only for a Chinese-language or region-specific context. Under the policy for natural-language violations, forcing a specific language without opt-in is a reportable issue.
This markdown file presents all user-facing content exclusively in Chinese, which can constitute a language/locale policy issue when no user opt-in or documented locale justification is provided. The file does not indicate that it is intended only for Chinese-speaking users or a region-specific audience.
The document title and all timeline content are written entirely in Chinese, which constitutes a language/locale constraint in the skill content. There is no indication that the user can choose another language or that the Chinese-only presentation is a documented, justified regional requirement.
This markdown file contains natural-language instructions and evaluation content exclusively in Chinese, and nowhere indicates that the skill is intended only for Chinese-speaking users or offers an alternative language. Under the language/locale policy rule, forcing a single language without opt-in can be a policy concern.