Security audit
Security checks for vulnerabilities and agentic risk
This skill is a coherent Xianyu listing helper that reads optional local reference files and a user-provided CSV, with no evidence of hidden exfiltration, persistence, destructive behavior, or credential use.
Installers should expect this skill to help draft and batch-normalize Xianyu listing materials. Only run the CSV helper on files you intentionally provide, and review generated pricing, condition disclosures, and service claims before posting them publicly.
Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.
Do not ask the user to perfect the data before starting. Fill gaps with assumptions and mark them.
## Output rules
### Titles
The skill instructs the agent to read local reference files and run a normalization script, but it declares no explicit tool scope or permissions. That mismatch can cause the runtime to grant broader-than-necessary file access or create ambiguity around what files may be read, increasing the risk of unintended local file exposure if the agent follows these instructions.
The activation text covers many common requests such as creating titles, descriptions, image prompts, reply scripts, and SOPs, with several 'also use when' clauses. Although Xianyu is named, the trigger scope is still broad enough that ordinary ecommerce copywriting requests could match unintentionally, and no exclusion examples or non-trigger cases are provided.
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.
Chinese headers like `商品名 / 类目 / 成色 / 价格 / 瑕疵 / 配件 / 城市 / 发货 / 备注` are also supported.
Do not ask the user to perfect the data before starting. Fill gaps with assumptions and mark them.
## Output rules
The file states 'Default output language: bilingual Chinese + English,' which imposes a language behavior by default rather than first asking the user's preference. The policy allows language constraints when the user is offered a choice or opts in, which is not stated here.
The template explicitly structures outputs into 'Chinese' and 'English' sections, which imposes specific language requirements in the skill content. There is no accompanying note that language should be selected based on user preference or context, so this can conflict with a language/locale choice policy.
The examples and later templates consistently frame outputs as Chinese and English, including a bilingual example and fixed Chinese/English title and description sections. This can amount to a language-policy issue because the skill appears to default to specific languages without explicitly offering the user a choice or documenting that the locale constraint is required.
The embedded prompt hint directs the downstream output to include 'CN titles' and 'a short CN description' as mandatory output components. This imposes a specific language choice in natural-language instructions without offering the user a language or locale option.
No suspicious patterns detected.