Back to skill

Security audit

直播回放爆点切片|AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a disclosed AI-HIVE video workflow, but it ships broader API client code and account/key handling that are not well aligned with the narrow livestream-clipping description.

Install only if you are comfortable using AI-HIVE for remote, potentially billable video generation and uploading user-selected media to that service. Prefer an environment variable or approved secret store for the API key, review commands before running them, and treat the bundled generic API client code as broader than the advertised livestream-clipping workflow.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • 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 (12)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

If the skill actually exposes generic chat, image generation, model enumeration, account information retrieval, wallet balance access, and browser-based API key setup while presenting itself as a narrowly scoped livestream clipping tool, users may be tricked into granting broader trust and providing credentials or sensitive account context under false pretenses. The danger comes from capability concealment and scope confusion: a disguised general-purpose/networked skill can access data or incur costs beyond what the user reasonably expects.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

If the skill actually exposes generic chat, image generation, model enumeration, account information retrieval, wallet balance access, and browser-based API key setup while presenting itself as a narrowly scoped livestream clipping tool, users may be tricked into granting broader trust and providing credentials or sensitive account context under false pretenses. The danger comes from capability concealment and scope confusion: a disguised general-purpose/networked skill can access data or incur costs beyond what the user reasonably expects.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

If the skill actually exposes generic chat, image generation, model enumeration, account information retrieval, wallet balance access, and browser-based API key setup while presenting itself as a narrowly scoped livestream clipping tool, users may be tricked into granting broader trust and providing credentials or sensitive account context under false pretenses. The danger comes from capability concealment and scope confusion: a disguised general-purpose/networked skill can access data or incur costs beyond what the user reasonably expects.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The file advertises and implements a generic AI API client with chat, image, video, upload, model discovery, and account-related operations, while the skill metadata claims a narrowly scoped livestream highlight clipping tool. This scope mismatch is dangerous because a host agent may grant or invoke the skill under the assumption of limited functionality, but the code enables broader capabilities that could be repurposed for unrelated content generation or data handling beyond the declared authorization boundary.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding

The skill advertises and demonstrates capabilities that imply access to environment variables, local files, shell commands, and network resources, but it does not declare any explicit tool scope or permission boundaries. In practice, this weakens least-privilege controls and makes it easier for a caller or future implementation to overreach into sensitive operations such as API-key handling, file modification, or outbound requests without clear authorization constraints.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The default prompt hard-codes Chinese output ('production-ready Chinese workflow'), which overrides user language preference without explicit opt-in. This is not a classic security exploit, but it can cause prompt steering and reduce user control, especially when downstream outputs must match policy, locale, or accessibility requirements.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/edit_video.py (reported line 20)May include surrounding context.

python
def run(args: list[str]) -> None:
    print(" ".join(args))
    subprocess.run(args, check=True)


def main() -> None:

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The documentation labels the code as a general-purpose AI tool and later as a raw-interface video skill, which conflicts with the manifest's narrowly defined livestream highlight intent. Misleading or contradictory documentation increases the chance that reviewers, orchestration systems, or end users will overtrust the skill's scope and permit broader actions than expected.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This code file contains natural-language descriptions, help text, and operational prompts exclusively in Chinese, beginning with the module docstring and continuing throughout the CLI UX. That can violate a language/locale policy when no user opt-in or alternative locale is offered.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The user-info capability exposes account and wallet information that is unrelated to the declared clipping workflow. In an agent environment, unnecessary access to billing or identity metadata expands the data exposure surface and could leak sensitive account details to users, logs, or downstream tooling without a legitimate business need.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The generic model-listing functionality allows enumeration of all available models, which exceeds the narrow purpose of a livestream clipping skill. This can reveal internal platform capabilities and make it easier to pivot into unintended use cases by discovering other text/image/video models not meant to be exposed through this specific skill.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
91% confidence
Finding

The file’s natural-language descriptions and skill identifiers are entirely in Chinese (for example the module docstring and argparse description), with no indication that users may choose another language or locale. Under the policy rule for language/locale constraints, this can be a natural-language policy issue because the skill appears to enforce a specific language without documented opt-in or justification.

Content

No source excerpt is available for this finding.