观众评论续写短剧｜AI-HIVE

• Outcome: Pass
• Latest audit: Aug 25, 2026 · 2:39 AM UTC
• Version: 1.0.0

Overview

This skill is a disclosed AI-HIVE media-generation workflow that uses user-provided API credentials and media files, with no artifact-backed evidence of hidden exfiltration or unsafe automatic behavior.

Install only if you are comfortable giving the skill access to an AI-HIVE API key and uploading the media files you explicitly pass to its commands. Review prompts and routing before paid generation, avoid placing API keys in logs or repositories, and use only media you are authorized to upload.

SkillSpector

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• 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
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (6)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 91% confidence
• Finding: The skill documents executable workflows that use environment variables, local file access, shell/ffmpeg, and networked API calls, yet it declares no permissions. That creates a transparency and consent gap: a user or hosting agent may invoke capabilities with greater reach than expected, including reading local files, writing outputs, and sending data to external services. In this context, the skill also handles media uploads and API keys, which increases the risk of unintended data exposure or unauthorized external actions.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 88% confidence
• Finding: The documented purpose is narrowly framed around comment-driven short-drama continuation, but the behavior appears to include broader capabilities such as generic AI chat, model/user-info querying, media upload utilities, and general local video manipulation. This mismatch is dangerous because it obscures the true attack surface and can mislead users, reviewers, or policy layers into approving a skill that can perform materially broader network, data-access, and content-processing actions than advertised. Because the extra features include external uploads and account-related queries, the hidden breadth meaningfully raises abuse and privacy risk.

Description-Behavior Mismatch

Medium

• Confidence: 95% confidence
• Finding: The file implements a broad, generic AI Hive client supporting chat, image, video, model listing, uploads, and task polling, while the skill metadata describes a narrower short-drama comment-driven image workflow. This capability mismatch increases attack surface and enables operators or downstream components to invoke unrelated functions not justified by the skill’s stated purpose.

Description-Behavior Mismatch

Low

• Confidence: 89% confidence
• Finding: Even the skill-specific CLI exposes generic `task` and `upload` operations that are not necessary for a narrowly scoped comment-to-image generation skill. While not inherently dangerous alone, they broaden what a caller can do with the user’s API key and can be combined with other features to step outside the intended workflow.

Description-Behavior Mismatch

Medium

• Confidence: 93% confidence
• Finding: The file presents itself as a narrow comment-driven short-drama skill, but it actually contains a broad API client that supports chat, model enumeration, account inspection, uploads, and arbitrary media generation. This capability mismatch expands the attack surface and enables uses outside the declared scope, undermining least privilege and making policy bypass easier in an agent setting.

Context-Inappropriate Capability

Medium

• Confidence: 89% confidence
• Finding: The skill can call the user-info endpoint and print account/profile and wallet-related data, which is unrelated to short-drama generation. In an agent environment, this creates an unnecessary data exposure path that could leak sensitive account metadata or balances to users, logs, or downstream tools.

Static analysis

No suspicious patterns detected.