Back to skill

Security audit

Seedream 5.0 Lite 广告图片

Security checks for vulnerabilities and agentic risk

Overview

The skill is a mostly coherent AI image-generation wrapper, but its result downloader can be steered by API response metadata to write outside the chosen output folder.

Install only if you trust AI Hive with your prompts, reference images, and generated outputs. Use a dedicated virtual environment, avoid sensitive campaign or customer data, choose a constrained output directory, and review or patch the downloader path handling before routine use.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/imagegen.py:350
Finding

Server-Controlled Download Metadata Allows Writes Outside the Output Directory

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
SKILL.md:74
Finding

Unpinned Runtime Dependency Installation Creates Supply-Chain Exposure

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (9)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The declared purpose emphasizes ad-to-landing-page consistency analysis, but the described behavior includes API-key initialization, local config writes, image upload, task polling, and result download, while the promised matching-chain validation is not actually implemented. This mismatch is dangerous because users may provide sensitive prompts or reference images under the assumption of a constrained auditing skill, when the real behavior sends data to an external service and performs broader side effects.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding

The skill exposes capabilities that imply shell, filesystem, environment, and network access, but it does not declare any tool scope or permissions boundary. That makes the operational trust model unclear and can lead users or hosting systems to invoke a skill with broader privileges than expected, increasing the chance of unintended file access, credential use, or outbound requests.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

This is a markdown file, so vague-trigger review applies. The description lists many broad phrases such as '广告图片', '效果广告', and '展示广告' without clearly constraining when this skill should activate versus other ad or image-related skills, increasing the chance of unintended invocation from common advertising-related requests.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The manifest description is written to direct usage in Chinese and does not offer a language choice or explain a justified locale restriction. Under the policy, forcing a specific language or locale without user opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill states that requests with a key are sent to a fixed external API and that results are downloaded, but it does not clearly warn that prompts and reference images will be transmitted off-platform. In an ad-creative workflow, prompts and images may contain unreleased campaign assets, customer data, or proprietary branding, so missing disclosure creates a real data-exposure risk.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest emphasizes building an information-matching chain between ad creatives and landing pages, including audience context, approved claims, and click-through consistency. In contrast, the implementation simply forwards prompt, reference images, and parameters to an image generation API and then polls/downloads results; there is no landing-page retrieval, claim verification, or consistency-checking logic.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The manifest describes a Seedream 5 Lite skill for creating ad images aligned with landing-page messaging, implying a focused generation workflow. The code also exposes separate task and upload subcommands that allow querying any generation task by ID and uploading reference images independently, which are broader utility capabilities not reflected in the manifest description.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The module docstring and all user-facing instructions are written only in Chinese, and the script presents prompts, errors, and guidance exclusively in that language. Under the policy, forcing a specific language without offering a user opt-in or alternative can be a natural-language policy violation.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

For an ad-image generation skill, interactive credential setup is understandable, but automatically launching a browser to a chat page with from=cli-skill and skill attribution parameters goes beyond the core image-generation purpose. This introduces a user-tracking or referral-oriented capability that is not disclosed in the manifest.

Content

No source excerpt is available for this finding.