Back to skill

Security audit

短剧场景连续性工作室|AI-HIVE

Security checks for vulnerabilities and agentic risk

Overview

This appears to be a real AI-HIVE media workflow skill, but it needs Review because its scripts can store an API key locally and send it to a user-configured API endpoint.

Install only if you trust this publisher and the AI-HIVE account boundary. Avoid using --base-url or AI_HIVE_BASE_URL unless you control and trust the endpoint, prefer a limited-scope API key, review ~/.ai-hive/config.json, and confirm parameters before any paid generation task.

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/imagegen.py:94
Finding

API Key Disclosure Through an Unrestricted Custom API Endpoint in Image Generation Client

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
scripts/videogen.py:94
Finding

API Key Disclosure Through an Unrestricted Custom API Endpoint in Video Generation Client

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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
Findings (20)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

A second, distinct mismatch is that the advertised workflow emphasizes structured continuity artifacts, while the exposed wrapper appears to center on general chat/media generation and local credential storage. When a skill presents itself as domain-specific but actually acts as a broader API client and content tool, users may unknowingly grant access to account data, keys, or generation functions outside the expected continuity task.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

A second, distinct mismatch is that the advertised workflow emphasizes structured continuity artifacts, while the exposed wrapper appears to center on general chat/media generation and local credential storage. When a skill presents itself as domain-specific but actually acts as a broader API client and content tool, users may unknowingly grant access to account data, keys, or generation functions outside the expected continuity task.

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 89)May include surrounding context.

md
python3 scripts/videogen.py init --skill-name short-drama-scene-continuity-ai-hive

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 91)May include surrounding context.

md
python3 scripts/videogen.py init --skill-name short-drama-scene-continuity-ai-hive

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding

The skill includes executable examples and implies capabilities involving environment variables, filesystem access, network calls, and shell execution, yet it does not declare any explicit tool scope or permissions. This creates an authorization boundary problem: an agent may expose or invoke broader capabilities than reviewers and users expect, increasing the chance of unintended file, network, or command execution.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
81% confidence
Finding

The description defines activation entirely through Chinese phrases and presents the skill as operating in a Chinese-language workflow, but it does not say this is optional or limited to a region-specific deployment. Under the language/locale policy, a skill should not impose a specific language without user opt-in or a clearly documented justification.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The invocation list includes a single generic term, '连续性', without narrowing context such as filmmaking, shot continuity, or scene management. Because this word is commonly used in many unrelated conversations, it could cause unintended activation of the skill.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The phrase '场景设定' can refer broadly to worldbuilding, creative writing, game design, or general scene setup, not specifically short-drama continuity work. The file does not provide negative examples or tighter constraints to distinguish intended use from other scene-design tasks.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The default prompt hard-codes Chinese output regardless of the user's language preference, which can override user intent and reduce transparency about what the skill is doing. While not a direct code-execution flaw, it can contribute to misleading behavior, review difficulty, and unsafe downstream use if users cannot easily verify generated workflows or commands.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill enables implicit invocation without any visible trigger constraints, so the platform may activate it in contexts the user did not clearly request. Because this skill is designed to generate production-ready workflows and runnable AI-HIVE commands, unintended invocation could cause unauthorized task execution paths, unexpected API usage, or user confusion about consent and scope.

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:

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The module docstring and all user-facing help and runtime messages in this code are written in Chinese, which imposes a specific language on users. The file does not offer an opt-in language selection or explain that the skill is intentionally region-specific, so this is a natural-language locale policy concern.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill is presented as a narrowly scoped short-drama scene continuity tool, but its wrapper forwards arbitrary prompts, images, batch size, routing mode, and free-form model parameters into a generic image generation backend. This scope mismatch weakens policy enforcement and enables use outside the declared purpose, including disallowed or higher-risk image generation workflows that downstream callers may wrongly assume are constrained.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The file is presented as a narrowly scoped short-drama scene continuity skill, but it actually ships a broad AI Hive client with unrelated capabilities including chat, image generation, account info retrieval, model enumeration, and generic media handling. This scope mismatch is dangerous because it increases the accessible attack surface and enables operators or downstream agents to invoke capabilities outside the declared purpose, undermining user consent and policy expectations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

This Python skill uses Chinese in its top-level description and throughout user-facing prompts, errors, and help text, which effectively forces a specific language on users. The file does not provide an opt-in language selection or explain that the skill is intentionally restricted to a Chinese-speaking or region-specific audience.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The generic text chat command is unrelated to a scene-continuity video tool and allows arbitrary prompts plus optional system prompts and image uploads. In a skill ecosystem, this creates covert capability expansion: a caller expecting constrained video functionality can instead use the skill as a general LLM interface, bypassing intended restrictions, auditing assumptions, or approval boundaries.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The script hard-codes Chinese-language descriptions and search aliases in user-facing strings, including the module docstring, CLI description, skill name, and alias list. Because there is no opt-in or alternative locale handling, this can violate language/locale policy requirements for user choice.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

Within the skill-specific CLI, the 'task' and 'upload' subcommands expose general-purpose AI-HIVE operational capabilities rather than continuity-focused scene-consistency logic. While media upload can support reference images, exposing standalone upload and raw task querying is infrastructure functionality not clearly justified by the manifest's end-user purpose as a short-drama scene continuity studio.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The manifest focuses on receiving scripts and references and delivering continuity artifacts or AI-HIVE-backed generation support. Fetching user profile and wallet balance is an account-management capability, not an obvious requirement for scene continuity work.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

While internal model lookup can be an implementation detail for generation routing, the public CLI command for listing all available models turns the skill into a general OpenAPI exploration tool. That capability is broader than the manifest's specialized continuity-studio framing.

Content

No source excerpt is available for this finding.