Security audit
Security checks for vulnerabilities and agentic risk
The skill mostly does what it claims, but it handles GeekBI login tokens in broader and less-disclosed ways than users should accept without review.
Install only if you are comfortable authenticating with GeekBI and with this skill creating local .geekbi/agent-auth.json state. Avoid using --base-url unless you have independently verified the endpoint is the intended HTTPS GeekBI service, and review or clear local auth files after use in shared, synced, or repository workspaces.
scripts/mercadolibre_review_search.py:55Unrestricted API Base URL Allows Authentication Data Transmission to Untrusted or Plaintext Origins
scripts/geekbi_auth.py:53Bearer Authentication State Is Replicated into Skill and Working Directories
The skill claims to retrieve and analyze product reviews by item or product ID, but the observed behavior instead queries site lists and resolves site metadata without performing the promised review retrieval. This mismatch is dangerous because operators may grant trust and execution privileges based on the stated purpose while the implementation does something materially different, undermining auditability and increasing the risk of hidden or substituted actions.
The skill claims to retrieve and analyze product reviews by item or product ID, but the observed behavior instead queries site lists and resolves site metadata without performing the promised review retrieval. This mismatch is dangerous because operators may grant trust and execution privileges based on the stated purpose while the implementation does something materially different, undermining auditability and increasing the risk of hidden or substituted actions.
The skill description is entirely written in Chinese and does not indicate that other languages are supported or that the language restriction is intentional for a region-specific use case. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.
The skill advertises executable behavior through referenced scripts and detected capabilities (file_read, network, shell) but does not declare any explicit tool restrictions. This creates an over-privileged execution surface where an agent may invoke filesystem, shell, or network actions without a clearly bounded scope, increasing the chance of unintended data access or command execution.
The display name, short description, and default prompt are all written in Chinese, which imposes a specific language/locale on users without indicating any option to choose another language. The file does not document that this skill is intentionally limited to Chinese-speaking users or a region-specific compliance context.
The file contains its docstring, CLI descriptions, and all user-visible error/action messages exclusively in Chinese, including messages emitted during login and storage failures. Because the skill does not offer a language/locale opt-in or indicate that it is restricted to a Chinese-speaking or region-specific environment, this is a natural-language locale policy concern.
The manifest describes a skill for querying and analyzing Mercado Libre reviews given a product ID, but this module implements a generalized GeekBI authentication/state subsystem with mirrored persistence locations, independent of any specific review-query logic. While network access to GeekBI is expected, maintaining reusable agent auth state across user config, skill directory, and working directory is a broader behavior than the manifest communicates.
The auth state is stored under a directory named 'temu-research-skill' even though this is a Mercado Libre review skill. That mismatch can cause unintended credential/state sharing or collision across different skills using the same local store, enabling one skill to read, reuse, overwrite, or clear another skill's login state on the same machine.
This Python file contains natural-language strings entirely in Chinese, including the module description and multiple raised error messages. Under the policy, forcing a specific language without user opt-in or a documented justification is a locale/language policy violation.
技能清单说明该技能用于基于明确的 Mercado Libre 商品 ID 查询和分析真实评论，但本文件实现的是 Mercado Libre 站点/国家/地区代码解析，并调用 /mercadolibre/site/ai-list 获取站点列表。站点发现能力并非“评论查询与分析”的直接实现细节，而是额外的数据解析功能，和清单描述的核心用途不一致。
This Python file contains its docstring, CLI description, help text, and error messages entirely in Chinese, which effectively constrains interaction to a specific language. The policy allows locale constraints only when users are given a choice or the constraint is clearly justified as region-specific, neither of which is present here.
This is a markdown file, so missing-warning review applies to user-facing behavior descriptions. Line L07 directs the system to re-run the original request after a minimal user reply, but the document does not disclose that the prior query will be repeated and may resend user or system data to the service.
The manifest presents a user-facing capability focused on retrieving and analyzing Mercado Libre product reviews, but this file also exposes a standalone command-line interface for inspecting storage and deleting auth state. Such local administrative functionality is not part of the stated review-search purpose and represents an extra capability beyond the user intent described in the manifest.
本脚本允许用户通过国家名、地区代码或站点 ID 查询并匹配 Mercado Libre 站点，这是一种目录/枚举型能力。对于一个宣称“必须先有明确商品 ID，只依据实际返回评论形成结论”的评论研究技能而言，该能力并无明显必要性，也未在技能描述中说明。