Back to skill

Security audit

children-science-video

Security checks for vulnerabilities and agentic risk

Overview

The skill is coherent for Chinese children's science-video production, but it needs review because it can auto-trigger a file-producing workflow from broadly related pasted text.

Install only if you want a Chinese-language children's science short-video workflow. Be careful pasting unrelated science or safety text after installation, because the skill tells the agent to take over automatically; confirm before any image or video generation that may consume credits, and review the generated output folder before using or publishing assets.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (15)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

声明描述的是一个覆盖文案生成、分镜设计、提示词优化、封面制作、社媒文案生成、图片/视频生成与文件归档的完整短视频生产技能;而实际代码仅是一个辅助质检脚本,检查分镜表中“配音文案”字段的长度是否符合规则。它既不处理用户原始科普文案,也不执行任何生成式内容生产、多媒体生成或自动触发逻辑。该代码的主要目的与声明的核心功能严重不符,因此属于明显的描述与行为不匹配。

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

声明描述的是一个覆盖内容创作、分镜设计、提示词生成、图片与视频生成、文件组织的完整短视频生产技能。而提供的代码只是单一的文档导出组件 make_platform_pdf.py:读取 Markdown 文件,使用 reportlab 和本地字体生成中文 PDF,并处理样式、emoji替换和页脚。它既不接收“任意科普文案”并产出脚本,也不调用任何图像/视频生成能力,更没有实现所谓自动触发的全流程。虽然声明中提到“PDF交付”,该代码最多只能算其中一个很小的导出环节,但无法代表所宣称的主要功能,因此描述与实际行为存在明显且重大的不匹配。

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The auto-trigger condition is extremely broad: nearly any pasted science, safety, or knowledge text causes the skill to take over and run a multi-step workflow. Overbroad triggers can cause prompt hijacking, unintended file/tool use, and surprise execution in contexts where the user did not request this skill, especially because the workflow includes reading files and generating artifacts.

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The instruction '任意科普文案即触发' lacks precise boundaries and says the trigger takes priority over everything else, so benign pasted content may unintentionally invoke the full skill. In an adversarial setting, a malicious user can embed qualifying text to force tool-selection and steer the agent into an unintended workflow, increasing the risk of prompt-routing abuse and unauthorized resource consumption.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
76% confidence
Finding

The skill instructs the agent to read multiple local files such as references/, assets/, and scripts/, but it does not declare any explicit tool scope or allowed-tools boundary. Missing scope declarations weaken least-privilege controls and can let the skill access more repository content than users expect if the runtime grants broad file-read capability.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

L027 规定“中文输出”,属于对语言/locale 的强制约束,但文档未说明这是用户可选项,也未给出区域合规或产品范围上的必要性说明。按规则,这构成自然语言层面的语言政策风险。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The file is entirely written as a prescriptive template in Chinese and includes hard requirements such as Chinese character counts in the voiceover rule, which effectively forces a specific language/locale. The policy allows locale constraints only when user choice is offered or the constraint is clearly justified as region-specific, neither of which appears here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The file is entirely written as a prescriptive standard for prompt construction in Chinese and repeatedly requires exact reuse of Chinese prompt blocks. There is no indication that users may opt into another language or that the Chinese-only requirement is justified by a documented region-specific constraint, which creates a natural-language locale policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The instructions explicitly require Chinese title text and discuss handling Chinese rendering issues, which indicates the skill is designed to force a specific language output. The file does not offer any user opt-in, alternative locale, or justification that this is a region-specific requirement.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The file's instructions are entirely prescriptive for Chinese-platform copywriting and specify Chinese stylistic conventions for titles, body text, and hashtags. There is no indication that users may opt into another language or that the restriction is explicitly justified as a region-specific compliance requirement.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

Multiple examples and hard requirements in the voiceover rules assume Chinese-only output, including explicit Chinese character count thresholds and prescribed Chinese phrasing. This creates a language policy issue because the skill appears to force a specific language without user opt-in or a documented region-specific justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The entire file prescribes prompt-writing rules in Chinese and includes fixed Chinese phrasing such as the required opening sentence and style constraints, but does not mention any user opt-in or alternative language option. Under the stated policy, forcing a specific language without choice is a natural-language policy violation unless the locale restriction is explicitly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The docstring states that each voiceover must have at least 36 Chinese characters and describes timing in terms of Chinese character count, which imposes a language-specific requirement. The runtime messages also reinforce this constraint, but the file does not offer any language choice or explain a justified region-specific limitation.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest describes an end-to-end儿童科普短视频 production skill that auto-generates scripts, storyboards, prompts, covers, social posts, images, videos, and stores all outputs in a titled folder. This file only converts an existing Markdown file into a PDF labeled as '双平台发布文案', which is a much narrower and different operation than the manifest's claimed behavior.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

该 markdown 文件通篇以中文撰写,并明确面向“抖音 / 小红书 / 视频号”等中文平台,包含固定的中文标题、文案和贴字内容。按规则,若技能在自然语言层面强制特定语言或 locale,且未提供用户 opt-in 或合理限定场景,属于语言/locale 政策风险。

Content

No source excerpt is available for this finding.