Security audit
Security checks for vulnerabilities and agentic risk
This is a disclosed Liuyao divination and learning skill with local reference files and helper scripts, and I found no hidden network access, credential use, automatic persistence, or destructive behavior.
Install only if you want a Chinese-style Liuyao divination and learning assistant. Treat readings as cultural or reflective content, not as medical, financial, legal, or major life decision advice; be aware that one optional maintenance script can append quality-check notes inside the skill's own logs directory if manually run.
The skill is presented as a divination/interpretation capability but includes maintenance behavior that reads local files and persists logs. That hidden operational behavior expands the attack surface and can expose internal content or create unauthorized state changes, especially when users invoke what appears to be a harmless end-user skill.
The skill is presented as a divination/interpretation capability but includes maintenance behavior that reads local files and persists logs. That hidden operational behavior expands the attack surface and can expose internal content or create unauthorized state changes, especially when users invoke what appears to be a harmless end-user skill.
The skill advertises executable/script-driven behavior and local file interactions, but does not declare any explicit tool or permission scope. In an agent environment, this can cause overbroad default access, making file reads/writes occur without clear operator review or least-privilege boundaries.
The instructions require the assistant to adopt a specific Chinese-speaking persona, including self-reference as "贫道" and addressing the user as "居士," with a fixed opening line. This imposes a language/style constraint without user opt-in, which is a natural-language locale policy concern under the stated rule.
The listed triggers such as "教我学六爻" and especially "帮我起一卦" are presented as direct activation phrases, but the file does not define boundaries, negative examples, or narrow activation context. This can create ambiguity about when the skill should activate versus when a user is making a casual or exploratory mention.
The skill states that this is an English adaptation and directs users to the Chinese version for fuller references, which imposes a language-specific constraint. The file does not present language selection or user opt-in, so this is a natural-language locale policy issue under the stated rules.
The phrase inviting users to say they want to consult Liuyao about a question is broad and natural-language-like, which can overlap with ordinary conversation and cause unintended skill activation. In agent environments where routing is based on loose semantic matching, this can trigger divination behavior when the user did not explicitly intend to invoke the skill, leading to confusion or policy-bypassing context switches.
The final 'Get Started' instruction uses an underspecified, conversational trigger that does not clearly define scope, exclusions, or invocation boundaries. This increases the chance of accidental activation from ordinary user statements about consulting Liuyao or discussing a question, especially in assistants that auto-select skills from free-form text.
This markdown log is entirely written in Chinese and repeatedly describes skill content and optimization goals in a way that suggests a Chinese-only operating mode. Under the policy rule, forcing a specific language without opt-in or documented regional justification is a natural-language policy concern.
This markdown file presents all instructions, templates, and examples exclusively in Chinese, and there is no indication that users can choose another language or that the skill is intentionally limited to a Chinese-speaking audience. The policy requires flagging language or locale constraints when they are imposed without user opt-in or explicit justification.
This section provides health-related divination about a father's illness improving and gives a time-based prognosis, but it does not clearly warn that the content is not medical advice and should not guide diagnosis or treatment decisions. Users may defer seeking professional care or change treatment behavior based on the prediction, which makes the omission safety-relevant even though the content is framed as divination.
The manifest describes a 六爻占卜 skill focused on generating divination readings and cultural interpretation, but this script performs repository-maintenance behavior by creating a logs directory and appending evolution records to a markdown file. That operational maintenance activity is outside the user-facing divination scope described in the manifest.
The code reads SKILL.md and reference/case markdown files, checks for persona slogans and specific philosophical references, counts examples, and generates editorial suggestions. This is a content-governance and documentation-quality workflow, which does not match the manifest's description of a divination skill that supports real/virtual casting and integrates Confucian/Taoist philosophy for readings.
The quality checks require specific Chinese-language cultural references and persona markers such as '贫道' and '居士', effectively enforcing one linguistic and stylistic locale. This is a natural-language policy issue because the script does not offer user opt-in or any documented justification for restricting outputs to that language/style.
This markdown file presents all instructions and content in a single language with no visible opt-in, alternative language option, or justification that the skill is region-specific. Under the policy, forcing a specific language without user choice can be a natural-language locale violation.
This markdown file presents all instructional and descriptive content in a single language, which can be a natural-language locale policy issue when no user opt-in or justification is provided. The file does not indicate that the skill is region-specific or that users may choose another language.
This markdown file presents all instructions and reference material exclusively in Chinese, with no note that the skill is Chinese-only or that users can choose another language. Under the language/locale policy, forcing a specific language without opt-in can be a natural-language policy violation.
This markdown file contains user-facing instructional content exclusively in Chinese, and there is no indication that the user can choose another language or that the locale restriction is intentional. Under the policy rule for language or locale constraints, forcing a specific language without opt-in can be a natural-language policy violation.
This markdown file presents all instructional content in Chinese and does not indicate that users can choose another language. Under the policy rule for language or locale constraints, forcing a specific language without opt-in is a natural-language policy concern.
The module docstring and user-facing prompts indicate the skill is intended to operate entirely in Chinese. Under the policy rule, forcing a specific language without user opt-in can be a natural-language policy violation when no choice or justification is provided.
The docstring says the function converts coin-toss results to 'binary', which suggests a straightforward binary encoding. In practice, it maps each three-coin result to a three-character pattern representing line types such as old/young yin-yang states, so the documentation is misleading about what the function actually returns.
The CLI instructions and validation logic require users to enter specific Chinese strings such as "正正反" and provide no language choice or opt-in. This enforces a locale-specific interaction model without documenting the restriction or offering alternatives.
This code file contains natural-language strings and documentation in a single enforced language, including the module description and later console output, but does not provide any opt-in or alternative locale handling. Under the stated policy, forcing a specific language without user choice is a natural-language policy violation.
No suspicious patterns detected.
