Security audit
Security checks for vulnerabilities and agentic risk
This is mostly a Chinese Markdown business guide, but it asks for shell access and advertises media-processing triggers that do not match its actual guidance-only capability.
Install only if you want a Chinese-language guide for planning video-editing monetization. Do not rely on it to process videos, edit audio, convert media, or generate dubbing, and prefer a version that removes exec permission and narrows its trigger text to business-planning use cases.
The trigger condition is overly broad and mismatched to the skill's real functionality, making the skill eligible for many unrelated media tasks. In an agentic system, broad misrouting is dangerous because it can systematically displace more appropriate skills, create faulty outputs, and normalize unsafe assumptions about what this tool can do.
The skill metadata and the entire user-facing content are presented only in Chinese, with no indication that language selection is optional or that the skill is intentionally limited to a Chinese-speaking or region-specific audience. This creates a natural-language locale constraint without explicit opt-in or documented justification.
The document repeatedly claims structured create/query/export operations using input_params, but the file is only a Markdown advice document and does not define executable interfaces. This mismatch can mislead an agent orchestration layer into treating unstructured content as an actionable tool contract, causing incorrect execution behavior or unsafe assumptions about available operations.
The skill is described as a monetization-planning guide, but its trigger conditions advertise media-processing capabilities such as video handling, audio editing, conversion, and voice generation. This can cause an agent to invoke the skill for tasks it cannot safely perform, producing misleading guidance or routing sensitive user requests into an unrelated workflow.
The trigger condition says the skill should be used for video processing, audio editing, media conversion, and dubbing generation, but the body only provides monetization advice. Overbroad and contradictory activation criteria increase the chance that an agent selects this skill for unrelated tasks, degrading reliability and potentially exposing user content to an unnecessary tool path.
The manifest grants exec tool access, while the availability section later says the skill is pure Markdown and does not require command execution. Unnecessary permission declarations violate least privilege and can increase the blast radius if an agent or wrapper trusts the manifest and enables shell access needlessly.
