Security audit
Security checks for vulnerabilities and agentic risk
This skill is mostly a disclosed marketing workflow, but it should be reviewed because it can automatically send business data to Feishu and implies ongoing profile learning without clear retention controls.
Install only if you are comfortable using a Chinese-language commerce assistant that may process product screenshots, campaign metrics, influencer details, and sales results. If enabling Feishu, use a dedicated webhook for the intended group, assume posted summaries may be visible to all group or document members, and disconnect it when no longer needed. Avoid relying on the implied preference-learning feature unless you know where the profile data is stored and how to delete or reset it.
The trigger phrases are very broad, such as generic requests for help with products or copywriting, which increases the chance the skill activates outside the user’s intended context. In an agent system, this can cause unintended collection, processing, or disclosure of business/product data and may override a more appropriate skill selection path.
The skill supports automatic pushing of analysis results, scripts, follow-up reminders, and review data to a Feishu webhook, but it does not present a clear warning that user-provided content will be sent to an external endpoint. This creates a real data-sharing risk because commercially sensitive product, campaign, and performance information may be transmitted to third-party chat spaces without sufficiently explicit informed consent.
The modular invocation table allows single-sentence requests like '帮我复盘' or '给我全案' to trigger substantial workflows without clear eligibility checks or precondition validation. This ambiguity can lead to accidental execution of sensitive business-analysis steps, especially when the skill auto-completes missing information or proceeds with incomplete context.
The file's instructions, templates, and headings are all presented only in Chinese, which imposes a specific language on users. Under the policy rule, this is a natural-language locale constraint unless the skill offers opt-in language selection or clearly documents a justified region-specific scope.
The skill describes automatically updating persistent user/category preference profiles from historical review data, but it does not mention notice, consent, retention limits, deletion controls, or how profiling data is stored and reused. In a commercial operations skill, this creates a real privacy risk because behavioral profiling can accumulate over time and influence future recommendations without the user's informed awareness or control.
The file states that selection analysis, reminders, follow-up prompts, and review reports can be automatically pushed to Feishu, but it does not adequately warn users about the content scope and possible exposure of business metrics, strategy details, or sensitive operational data to all members of the destination chat or document. In this skill context, outbound sharing is a core feature, which increases the chance that commercially sensitive information is distributed more broadly than the user expects.
The document defines multiple automatic Feishu push triggers in broad terms without explicit scoping, confirmation, or per-event opt-in. This can cause analysis results, operational plans, or follow-up content to be sent to a configured webhook when the user did not clearly intend outbound delivery for that specific run, creating a real risk of unintended disclosure to a group.
Using a natural-language phrase like “帮我安排推品” as a trigger is ambiguous and lacks safeguards around scope, timing, and target campaign. In a conversational system, such loose triggering can misfire from casual discussion or partial context and initiate external notifications to Feishu without sufficiently specific user authorization.
The file includes ready-to-use influencer outreach and follow-up scripts that encourage pressure tactics such as scarcity, social proof, and competitive urgency without any compliance guardrails. In a commercial agent that automates end-to-end promotion, this can facilitate deceptive or non-compliant outreach, increasing legal, platform-policy, and reputational risk at scale.
The document title and all operative instructions are written exclusively in Chinese, and the file does not indicate that language selection is optional or limited to a region-specific use case. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.
The file content is entirely in Chinese and presents fixed output templates without any indication that language should follow the user's preference, which can force responses into a specific language without opt-in. While not a direct code-execution risk, this can degrade safety and usability by obscuring important compliance or operational content for users expecting another language, and can contribute to misrouting or misunderstanding in multilingual deployments.
The default trigger phrase is effectively the broad user utterance '帮我出素材', which is common, unspecific language that could easily be invoked unintentionally in adjacent contexts. In an agent skill system, overly broad activation increases the chance of accidental routing into this skill, causing unsolicited generation of marketing assets and possible context hijacking away from the user's actual intent.
The file's headings, instructions, and examples are all in Chinese, and there is no natural-language indication that users may choose another language or that the skill is intentionally limited to a Chinese-language or region-specific workflow. The policy specifically calls for flagging language or locale constraints when they are imposed without user opt-in or clear justification.
The instruction "所有输出使用简体中文" imposes a fixed language policy for all outputs. The file does not present this as an opt-in preference or a justified region-specific constraint, so it conflicts with the requirement to avoid forced language settings without user choice.
SQP-3 applies to all file types and covers language or locale policy violations. This skill content is entirely in Chinese and does not mention alternative language support or that Chinese is an intentional, user-selected locale, so it appears to impose a language choice without opt-in.
The skill content, examples, and instructions are all presented in Chinese, which can amount to a language policy constraint if users are not given an opt-in or alternative locale. The file does not state that the skill is intentionally limited to Chinese-language users or a China-specific compliance context.
No suspicious patterns detected.
