Security audit
Security checks for vulnerabilities and agentic risk
This contract-drafting skill is coherent, but it should be reviewed because confidential contract text may be sent to an external API under unclear privacy disclosures.
Install only if you are comfortable sending contract drafts and related commercial details to the Yunlv TradeGPT API. Redact bank account details, personal data, and unnecessary confidential terms where possible, and treat generated contracts as drafts that still need qualified legal review.
The skill states that contract content is 'not stored on third-party servers' while also explicitly allowing calls to an external TradeGPT API for contract generation and analysis. That creates a misleading privacy representation around highly sensitive commercial and legal data, which can cause users to disclose confidential contracts under false assumptions about data handling.
Using the generic trigger term 'contract' can activate this skill for many unrelated requests, causing accidental invocation in contexts where users did not intend to share legal or commercially sensitive material. In a skill that may transmit full contract text to an external API, unintended activation increases the chance of privacy leaks and inappropriate handling of user data.
The skill permits sending contract content to an external API but does not place a clear, user-facing privacy warning near the usage flow where users are instructed to paste contract text. Because contracts often contain confidential pricing, counterparties, banking details, and dispute terms, omission of an immediate disclosure undermines informed consent and can lead to unintentional data exposure.
The manifest description explains the skill's drafting and review functions but does not disclose that user-supplied contract content may be transmitted to the external TradeGPT API at api.yunlvai.com. In a contract-review context, inputs may contain confidential commercial terms, pricing, counterparties, and legal obligations, so lack of notice undermines informed consent and creates privacy and confidentiality risk.
The manifest uses broad English and Chinese trigger terms such as "contract", "NDA", and "Sales Contract", which can cause the skill to activate for many ordinary legal or business conversations. Because this skill sends content to an external API for contract generation and risk analysis, unintended invocation increases the chance that sensitive contract drafts or negotiation details are disclosed to a third party without clear user expectation.
The language-selection table prescribes recommended contract languages based on geography, such as '中英双语' for multiple regions, rather than presenting language choice as user-directed. This can be a natural-language policy violation because it imposes locale preferences without explicit user opt-in or a clearly justified compliance constraint.
The entire skill content is presented only in Chinese, beginning with the title and continuing through all instructional/reference material. Under the policy, a language-specific constraint should either provide user choice/opt-in or be clearly documented as a justified region-specific limitation, which is not present here.
SQP-3 applies to all file types and covers language or locale policy violations in natural-language strings. Here, the module docstring and command descriptions are presented only in Chinese, and the file does not offer an opt-in language selection or explain that the tool is intentionally region-specific.
Overly Broad Trigger: 'PI' is too short and may match unintended inputs
This markdown file contains user-facing natural language content exclusively in Chinese, beginning with the title on L01, and nowhere indicates an opt-in language preference or a justified locale restriction. Under the policy rule, forcing a specific language without user choice can be a natural-language policy violation.
Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.
Force Majeure events shall include but not limited to:
- Natural disasters (earthquakes, typhoons, floods)
- War, terrorism, civil unrest
- Government actions, embargoes, sanctions
Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.
Force Majeure events shall include but not limited to:
- Natural disasters (earthquakes, typhoons, floods)
- War, terrorism, civil unrest
- Government actions, embargoes, sanctions
No suspicious patterns detected.