Reels Script Studio

• Outcome: Pass
• Latest audit: Sep 16, 2026 · 10:17 AM UTC
• Version: 3.5.0

Overview

This is a coherent short-video script generation skill with a disclosed local learning file; it does not show exfiltration, hidden execution, or destructive behavior.

Install only if you are comfortable with the skill saving local usage statistics, preferences, and feedback in evolution.json. Avoid recording sensitive campaign details, private client data, or unreleased product information unless you are willing for that information to remain in the local skill folder.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (14)

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 96% confidence

Finding

The declared description promises a feature-rich generator whose primary function is producing 15-second viral short-video scripts and related prompts. The supplied code chunk does not implement any generation behavior at all. Instead, it is an auxiliary evolution/telemetry utility that persists usage records and feedback to evolution.json, increments a version number, and prints reports. While this does align with the narrowly mentioned 'self-evolution mechanism (evolution.json),' that is only one supporting aspect of the declared skill, not its main purpose. Therefore, the code chunk materially differs from the declared functionality and should be flagged as a mismatch.

Content

No source excerpt is available for this finding.

Ae1

High

• Category: analysis-evasion
• Confidence: 100% confidence

Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 129)May include surrounding context.

md

核心文件（`SKILL.md`、模板、运镜库、爆款手册、质量校验）为**冻结层**，禁止在进化中改动结构，仅可调整参数值。

Undeclared Tool Scope

Medium

• Category: MCP Least Privilege
• Confidence: 84% confidence

Finding

The skill documents persistent state changes via evolution.json and scripts/evolve.py, implying read/write behavior, but it does not declare any explicit tool scope or permissions. Undeclared file access weakens least-privilege controls and makes it harder for operators and users to understand that the skill may persist data or modify local files.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

The trigger phrases include broad, common requests such as asking for a script or prompt help, which can cause the skill to activate in contexts the user did not intend. Over-triggering can lead to unexpected behavior, unsolicited collection into evolution.json, or interference with other skills and workflows.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

• Category: Not specified by scanner
• Confidence: 94% confidence

Finding

The skill is presented primarily as a script generator, but it also advertises persistent self-evolution, usage statistics, and user preference storage in evolution.json. Silent profile accumulation creates privacy and security risk because users may not expect long-term retention of their interactions or inferred preferences from a content-generation skill.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 98% confidence

Finding

The self-evolution feature stores usage and preference data, but the user-facing description does not prominently warn about this behavior. Lack of upfront disclosure undermines informed consent and can result in unexpected retention of personal or behavioral data entered during script creation.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

• Category: Not specified by scanner
• Confidence: 90% confidence

Finding

The documented analytics and profile accumulation go beyond what is strictly necessary to generate a one-off script. Collecting usage_stats, user_preferences, and experience_log without strong purpose limitation increases exposure of behavioral data and expands the consequences of unauthorized file access or misuse.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 94% confidence

Finding

The value "Markdown完整版" sets a specific language/locale preference directly in the configuration. Under the policy, forcing a specific language without offering user choice or documenting a justified regional constraint is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 91% confidence

Finding

The file hard-codes Chinese city/localized hashtag guidance such as #北京美食 and #上海打工人的夜 without first checking the user's language, region, or target audience. In a multi-platform script-generation skill, this can cause unintended locale inference, inappropriate geographic targeting, and reduced user control over how their content is localized or presented.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

The entire template file is written in Chinese and presents fixed Chinese script patterns and phrasing for use, without indicating that users may choose another language or locale. This can violate language/locale policy when a skill constrains outputs to a specific language without opt-in or documented regional justification.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium

• Category: Not specified by scanner
• Confidence: 94% confidence

Finding

The guide says users can input phrases like "帮我写一个15秒短视频脚本" / "write me a 15s short-video script," which are generic requests that could overlap with normal conversation rather than a clearly bounded invocation syntax. It also presents them as examples ("等触发句" / "a trigger like") without listing exact allowed triggers or exclusion conditions, so the skill's activation boundary is unclear.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 93% confidence

Finding

The guide states that user-derived preference/history data is persisted to a local evolution.json file, but it does not prominently warn users that data retention and file creation will occur. Persistent storage of interaction-derived data can expose sensitive business prompts, creative preferences, or personal information to other local users, backups, sync tools, or downstream handling without informed consent.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 91% confidence

Finding

This file presents all instructional content in a single language and does not indicate that users may opt into Chinese or switch to another language. Under the policy, forcing a specific language without user opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

• Category: Not specified by scanner
• Confidence: 80% confidence

Finding

The Chinese quick-start says users should reply with menu codes, implying a fixed Chinese-language interaction pattern at launch. Although the document is bilingual overall, it does not explicitly state that users may choose English instead of the Chinese interaction flow, so the language choice is not clearly offered as an opt-in.

Content

No source excerpt is available for this finding.