Back to skill

Security audit

Feed Designs You Video Gen

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed WeryAI video-generation helper for a specific short-video template, with no evidence of hidden persistence, local data collection, or destructive behavior.

Install only if you are comfortable using a paid WeryAI API key and sending the video prompt plus any supplied image URLs to WeryAI. Use public HTTPS image URLs only, avoid sensitive or private content, and confirm the expanded prompt before generation because the local script does not itself enforce every documented image URL restriction.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (1)

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/video_gen.js:188
Finding

Unvalidated Image URLs Forwarded to a Remote Generation Service

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
Findings (9)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The declared description presents a narrowly tailored 'algorithm-control shorts' generator, but the body exposes a more general remote video-generation workflow with multiple modes, model selection, and operational API usage. This mismatch can mislead users and policy layers about what the skill is actually capable of, which weakens review, consent, and least-privilege controls.

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 34)May include surrounding context.

md
- **API hosts** fixed in `video_gen.js`.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding

The skill declares network use and access to a secret environment variable but does not define an explicit tool scope such as permissions or allowed-tools. That creates an authorization gap where the agent may invoke broader capabilities than users expect, increasing the risk of unintended external calls or secret-adjacent operations.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The description includes broad trigger phrasing such as 'Use when the user says the app chooses for them,' which can cause overbroad or unintended activation. Over-triggering is dangerous because it may route ordinary user requests into a networked skill that transmits content externally without sufficiently specific user intent.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill states that it uses WERYAI_API_KEY and networked video generation but does not provide a clear user-facing disclosure that prompts, images, and other content will be sent to an external service. This creates a data-handling and consent risk, especially if users provide sensitive text or images assuming local processing.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The implicit trigger prompts are broad enough that ordinary requests about scrolling apps, recommendations, UI montages, or captioned vertical videos could activate the skill even when the user did not explicitly intend to use it. In an agent setting, this can cause over-triggering, unintended prompt expansion, and submission workflows to run in contexts outside the narrow 'algorithm-control critique' use case, increasing the chance of unauthorized or confusing behavior.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest describes a specialized skill for creating vertical shorts with a specific rhetorical structure, but this file documents a broad-purpose video generation interface that fetches arbitrary model metadata and explicitly leaves per-skill constraints unenforced. The note that callers must supply the model required by the skill shows the underlying behavior is not limited to the manifested niche behavior.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The manifest focuses on creating captioned vertical montage shorts for feed/algorithm rhetoric, but the file advertises separate image-to-video and multi-image-to-video modes as first-class operations. Those broader generation modes materially expand the skill beyond the described specialized short-form template.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The top-level docstring presents the CLI description, commands, examples, and environment-variable guidance entirely in Chinese. For a general-purpose CLI, this imposes a specific language on users without any opt-in, fallback, or stated region-specific requirement, which matches the locale/language policy violation criteria.

Content

No source excerpt is available for this finding.