Back to skill

Security audit

AI大模型专家|AI公司AI图片视频营销

Security checks for vulnerabilities and agentic risk

Overview

This skill mostly matches its stated AI-HIVE marketing purpose, but a configurable server address could expose your API key and uploaded media if changed.

Install only if you are comfortable using AI-HIVE with an API key and uploading selected campaign media. Keep the default AI-HIVE endpoint, avoid AI_HIVE_BASE_URL or --base-url unless you control the server, and use a limited API key because a changed endpoint could receive your key and media.

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:91
Finding

Unvalidated API Endpoint Override Can Expose API Credentials and User Media

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
scripts/videogen.py:91
Finding

Unvalidated Video API Endpoint Override Can Expose API Credentials and Uploaded Media

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 (17)

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The documentation presents a broad marketing/content-production workflow, but the code reportedly includes undeclared user-info, wallet, chat, and image-generation functionality while lacking the advertised persistence and reporting features. This is dangerous because users may share business data expecting a scoped marketing assistant, yet the skill may access adjacent account features or perform chargeable generation outside the described contract.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The documentation presents a broad marketing/content-production workflow, but the code reportedly includes undeclared user-info, wallet, chat, and image-generation functionality while lacking the advertised persistence and reporting features. This is dangerous because users may share business data expecting a scoped marketing assistant, yet the skill may access adjacent account features or perform chargeable generation outside the described contract.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The documentation presents a broad marketing/content-production workflow, but the code reportedly includes undeclared user-info, wallet, chat, and image-generation functionality while lacking the advertised persistence and reporting features. This is dangerous because users may share business data expecting a scoped marketing assistant, yet the skill may access adjacent account features or perform chargeable generation outside the described contract.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding

The documentation presents a broad marketing/content-production workflow, but the code reportedly includes undeclared user-info, wallet, chat, and image-generation functionality while lacking the advertised persistence and reporting features. This is dangerous because users may share business data expecting a scoped marketing assistant, yet the skill may access adjacent account features or perform chargeable generation outside the described contract.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding

The skill advertises executable code paths involving shell commands, file access, environment variables, and network use, but declares no explicit tool scope or permission boundaries. In an agent environment, this can cause the skill to run with broader ambient authority than users or reviewers expect, increasing the chance of secret exposure, unintended file modification, or unauthorized outbound requests.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The description says to use the skill when a user searches or asks about a long list of generic terms such as AI图片、图片生成、视频生成、内容营销、品牌营销 and 获客推广. These phrases are broad and overlap with many ordinary requests, while the file does not clearly define exclusion boundaries beyond a later 'not suitable' section, increasing the risk of unintended invocation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The skill content is entirely presented in Chinese and positions itself for use based on Chinese-language search or query patterns, but it does not state that language selection is optional or user-driven. Under the policy, forcing a specific language or locale without opt-in can be a natural-language policy violation unless the restriction is explicitly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill metadata and default prompt enforce Chinese output regardless of the user's language preference, which can override user intent and cause misleading or unusable responses. While this is not a direct code-execution issue, it is a prompt-safety and user-consent problem that can degrade reliability, conceal important warnings, or cause the system to behave unexpectedly in multilingual contexts.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The invocation scope is broad and loosely constrained: the default prompt and implicit invocation settings make it easy for the skill to activate on general AI marketing or media-generation queries without clear user intent boundaries. In a skill that also promises AI-HIVE code and task routing behavior, ambiguous triggering increases the risk of unintended execution paths, over-collection of user inputs, or surprise automation the user did not explicitly request.

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

This Python skill presents its top-level description and most CLI guidance in Chinese, and there is no mechanism for the user to select a preferred language or opt in to this locale. The policy scope explicitly covers natural-language strings in code files, so a hard-coded single-language experience is a locale-policy issue here.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill embeds core user-facing content and defaults entirely in Chinese, including industry, mode, focus, channels, and boundary text. This enforces a specific language/locale without offering user opt-in or documenting that the skill is intended only for a Chinese-language or region-specific workflow.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This code file contains natural-language descriptions, help text, and user-facing interaction entirely in Chinese, beginning with the module docstring and continuing throughout the CLI experience. Under the policy rule, forcing a specific language or locale without user opt-in is a natural-language policy violation unless the constraint is explicitly documented and justified, which is not present here.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest describes this skill as a marketing-oriented AI-HIVE tool for current image/video models, asset upload, pricing snapshots, routing, polling, and task tracking. However, this file includes a full generic chat command that can send arbitrary text and multimodal chat requests unrelated to the declared image/video marketing scope, broadening behavior beyond what the manifest claims for this skill.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
23% confidence
Finding

The file-level docstring says the brief is created 'without network calls', but this script is only a local JSON generator and does not cover the networked AI-HIVE querying, uploading, routing, polling, or task recording described in the manifest. This is more of a narrowing/incompleteness issue than a direct contradiction, so confidence is low and it is not a strong intent-divergence finding.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The manifest emphasizes delivering marketing strategies and image/video generation workflows via AI-HIVE. In contrast, the file includes generic user-info and raw models inspection commands that retrieve wallet/user information and full model listings, which are broader platform utility operations rather than the declared marketing skill behavior.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The manifest describes content strategy, generation, uploads, pricing snapshots, routing, polling, and task recording. The init flow additionally launches a browser to a URL containing from=cli-skill and skill=<name> for backend attribution, which is ancillary telemetry-style behavior not necessary to perform the declared marketing generation functions.

Content

No source excerpt is available for this finding.