ClawHub
Back to skill

Security audit

开放世界都市任务感游戏预告｜AI-HIVE原创工作流

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly fits its AI-HIVE video-generation purpose, but it needs Review because it can send credentials to an endpoint controlled by an environment variable and can be invoked broadly for paid-capable external generation.

Review before installing. Prefer OAuth or a managed secret store, avoid setting AI_HIVE_MCP_URL unless it is exactly the intended AI-HIVE endpoint, and revoke any AI-HIVE key that may have been exposed. Be aware that the skill may be selected for broad creative requests, and require clear confirmation before uploads, paid generation, batching, sending, or publishing.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (4)

Tainted flow: 'request' from os.environ.get (line 67, credential/environment) → urllib.request.urlopen (network output)

Critical
Category
Data Flow
Content
method="POST",
    )
    try:
        with urllib.request.urlopen(request, timeout=60) as response:
            result = parse_payload(response.read(), response.headers.get("content-type", ""))
            return result, response.headers.get("mcp-session-id") or session_id
    except urllib.error.HTTPError as error:
Confidence
94% confidence
Finding
The script sends credentials from environment variables to a remote endpoint whose base URL is taken from AI_HIVE_MCP_URL without any allowlist or origin validation. If an attacker can influence that environment variable, they can redirect requests and exfiltrate the API key or bearer token to an arbitrary server.

Lp3

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding
The skill instructs users to use networked MCP endpoints, local scripts, environment variables, and local file output, which are operational capabilities with security implications, but it does not declare permissions accordingly. This can bypass user/admin expectations about what the skill may access and increases the risk of unintended network calls, local file writes, or secret handling in less controlled clients.

Vague Triggers

Medium
Confidence
78% confidence
Finding
The description contains broad search-style triggers such as GTA 6 style demos, anime/game, AI image/video, and short-video production, which can cause the skill to activate for many general creative requests beyond its narrowly described workflow. Over-broad routing can expose users to unnecessary external-tool usage, data collection, or paid-generation prompts when they did not intend to invoke this specific skill.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The skill enables implicit invocation without defining narrow trigger constraints, so the agent may auto-select this skill in loosely related user requests. Because the skill connects to an external MCP capable of content generation and potentially paid operations, unintended invocation could cause unreviewed third-party data sharing, unexpected actions, or user confusion about when the tool is being used.

Static analysis

No suspicious patterns detected.