Back to skill

Security audit

Meme梗图短视频生成器|AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

This looks like a legitimate AI-HIVE meme-video helper, but it has under-scoped credential and file-write risks users should review before installing.

Install only if you trust the publisher and are willing to review the helper script before use. Keep the API host at the official AI-HIVE URL, do not use untrusted --base-url or AI_HIVE_BASE_URL values, and rotate the API key if it may have been used with a custom endpoint. Treat generated downloads as untrusted files and confirm all paid generation parameters before running commands.

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 (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/videogen.py:92
Finding

Bearer API Credential Disclosure Through Unrestricted Base URL Configuration

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/videogen.py:373
Finding

Arbitrary File Write Through Unsanitized API Task Metadata

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (13)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The declared description presents a full meme short-video generator and AI-HIVE-integrated workflow for producing rewritten branded meme content and managing media-generation jobs. The supplied code does not perform any of those core functions. It only parses command-line arguments and writes a static JSON brief with metadata and checklist-style fields. While this could be a small supporting component of the broader skill, the chunk’s actual primary behavior is much narrower and materially different from the declared end-to-end generation and API orchestration capabilities.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The declared description presents a higher-level meme/brand-content generation skill with external API-driven media generation and workflow orchestration. The supplied code instead implements a simple local CLI wrapper around ffmpeg/ffprobe for basic editing tasks. While short-video editing could be a supporting detail for such a skill, the code lacks the central described behaviors and integrations, and its actual primary purpose is materially different from the declared purpose.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
89% confidence
Finding

The declared description presents a domain-specific creative skill centered on meme remixing for brands and social media operations, including higher-level editorial outputs like explaining meme structure, rewriting it into an original branded expression, and producing subtitles/publishing variants. The code, however, mostly implements a reusable AI Hive command-line SDK: it can initialize API keys, open a browser, query user info and wallet balance, list/find models, upload media, invoke generic text/image/video endpoints, poll tasks, and download assets. While the wrapper at the bottom configures preset video models and supports the AI-HIVE operational details described (pricing snapshot, routing modes, upload, polling, download), it does not encode the specialized meme-analysis or brand-content transformation workflow described as the main purpose. It also exposes extra capabilities not declared, notably generic chat, image generation, model enumeration, and user-info access. Therefore the description does not accurately represent the actual code's behavior.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill embeds executable command examples and describes capabilities involving environment variables, file access, network access, and shell usage, but it declares no explicit tool scope or allowed-tools boundary. In an agent setting, that omission can cause overbroad tool access, making it easier for the skill to trigger network calls, read/write local files, or expose secrets without clear authorization constraints.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The entire skill description, triggers, outputs, and examples are written as Chinese-only instructions, with no indication that users may choose another language or that the skill is intentionally restricted to a Chinese-language or region-specific context. Under the stated policy, forcing a specific language without user opt-in can be a natural-language policy violation.

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 terms such as social media content and brand rejuvenation concepts, which can cause the skill to activate in unrelated conversations. In an agent environment, unintended invocation can route user data into the wrong workflow, increase exposure to networked or billable actions, and produce actions the user did not intend.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Line L4 instructs the skill to produce a "production-ready Chinese workflow," which imposes a language choice in the default behavior. The file does not indicate that the user can choose another language or opt into Chinese, so this appears to violate language/locale choice expectations.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest sets allow_implicit_invocation to true without any visible trigger constraints or narrowing conditions, so the skill may be auto-selected for loosely related requests. Because this skill can generate runnable workflows and AI-HIVE commands for media generation, ambiguous activation increases the chance of unintended execution paths, misuse for policy-sensitive content, or user confusion about what tool is acting.

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:

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The file presents itself as a generic AI Hive OpenAPI client with chat, image, video, model enumeration, upload, and polling features, which materially exceeds the manifest’s stated purpose of meme short-video generation. Over-broad capabilities increase attack surface and enable callers to use the skill for unrelated actions, undermining least privilege and creating scope-creep risk if this skill is trusted as narrowly scoped.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The general-purpose text chat function allows arbitrary model interaction beyond the advertised meme video generator behavior. In skill ecosystems, such hidden or unnecessary capabilities can be abused to bypass policy/scoping controls, process unrelated data, or repurpose the skill as a broad AI proxy instead of a narrow media-generation tool.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

Standalone image generation is broader than the manifest’s specific meme short-video generator purpose and creates an unnecessary extra content-generation surface. This mismatch makes the skill more dangerous because users or downstream agents may trust it as narrowly scoped while it can actually perform unrelated media generation tasks.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill includes a user-info endpoint that retrieves account and wallet balance data, which is unrelated to generating meme short videos. While not directly destructive, it expands the skill’s access to potentially sensitive billing/account metadata and violates least-privilege expectations for a content-generation skill.

Content

No source excerpt is available for this finding.