Security audit
Security checks for vulnerabilities and agentic risk
This is a disclosed Chinese divination skill made of Markdown instructions and examples, with no executable code or hidden high-impact behavior.
Install only if you want a Chinese-language Xiaoliuren divination assistant. Treat its readings as entertainment or cultural reference, not as a substitute for medical, legal, financial, or other important decisions.
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
git clone https://github.com/tianmingwangluo/xiaoliuren-skill.git
mkdir -p ~/.dsh/skills/xiaoliuren-skill
cp xiaoliuren-skill/SKILL.md xiaoliuren-skill/examples.md ~/.dsh/skills/xiaoliuren-skill/
The README states that ordinary natural-language requests like asking for divination will automatically load the skill. This trigger is broad and overlaps with common user speech, which can cause unintended invocation and make prompt-routing easier to manipulate by a user or another skill. In a skill system, overly permissive activation conditions increase the chance the skill is pulled into contexts where it was not explicitly requested.
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.
git clone https://github.com/tianmingwangluo/xiaoliuren-skill.git
mkdir -p "${HERMES_HOME:-~/.hermes}/skills/xiaoliuren-divination"
cp xiaoliuren-skill/SKILL.md xiaoliuren-skill/examples.md "${HERMES_HOME:-~/.hermes}/skills/xiaoliuren-divination/"
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
也可从 GitHub 直接安装（会经过 Hermes 安全扫描；仅下载 SKILL.md，不含 examples.md）：
hermes skills install https://raw.githubusercontent.com/tianmingwangluo/xiaoliuren-skill/main/SKILL.md
安装后：
The Hermes section says the platform will auto-load the skill when a user describes a divination request, but it does not define constraints, exclusions, or disambiguation rules. That ambiguity can lead to accidental activation on loosely related messages and creates a wider surface for routing abuse or unintended prompt injection through skill activation. Because this is a divination skill, broad loading is unnecessary and should be tightly scoped.
The general guidance for other AI agents says they will automatically call the skill whenever the user requests xiaoliuren divination. This is an overbroad claim that encourages deployers to configure automatic routing without guardrails, increasing the likelihood of misfires and untrusted-context activation. In multi-skill environments, this can cause the skill to intercept routine conversation or be invoked by crafted phrasing.
The instruction says the interpretation must be delivered in '口语化中文', which forces a specific language for responses. This is a natural-language policy concern because the file does not offer the user a language choice or indicate that Chinese is optional.
The provided response template is entirely in Chinese and is presented as the required output format, reinforcing a Chinese-only interaction model. Because the skill is marked user-invocable and does not provide any language-selection mechanism, this can violate language/locale choice expectations.
This file’s natural-language instructions and examples are presented only in Chinese, and there is no indication that the user can choose another language or has opted into this locale. Under the policy, forcing a specific language without user opt-in can be a natural-language policy violation.
No suspicious patterns detected.