Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed daily Chinese poetry sender that reads bundled poem data and posts generated content to Feishu, with some quality and consent caveats but no hidden or malicious behavior found.
Install this only if you want a scheduled Chinese-language poetry message sent to Feishu. Confirm the Feishu destination and schedule are acceptable, and be aware the dataset should be corrected because it includes at least one non-Tang poem and a mismatched poem count.
The skill invokes local resources (data/tang-poems.json via a Python script) but does not declare any explicit tool scope or permissions boundary. Even though the documented use is narrow and benign, undeclared file-read capability weakens reviewability and can allow future modifications to read unintended files without users or policy controls clearly seeing that access.
The trigger phrases are broad enough that normal conversation about Tang poetry could unintentionally activate the skill, causing unexpected automated behavior. In this skill's context, accidental activation is more concerning because the workflow culminates in sending a message to Feishu, which could produce unsolicited outbound content.
The skill instructs the agent to send generated content to Feishu without clearly warning the user that content will be transmitted to an external messaging channel. This creates a transparency and consent problem: users may believe they are only generating text locally, while the skill actually performs an outbound delivery action.
This JSON dataset stores poem titles, content, and author descriptions exclusively in Chinese throughout the file. For a general-purpose skill asset, that constitutes a natural-language locale constraint with no indication that users can choose another language or that the Chinese-only scope is explicitly justified.
The dataset contains at least one poem explicitly marked as Song dynasty ("宋") while the skill claims to provide Tang poems from the Complete Tang Poems corpus. This is a data integrity issue: users may receive incorrect content, which undermines trust and can propagate factual inaccuracies in downstream outputs or educational use.
Natural-language instructions describe the skill entirely in Chinese and specify Chinese-formatted output for the generated push content, but do not indicate that language is configurable or that the user can opt into Chinese output. This can violate language/locale policy when a skill implicitly forces one language without user choice.
The docstring and generated date format explicitly require Chinese output context (for example, "每日唐诗" and "%Y年%m月%d日") with no indication that the user can opt into another language or locale. This is a natural-language locale constraint that appears hard-coded rather than user-selectable.
