Back to skill

Security audit

京东 电商图片生成与编辑

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent AI Hive image-generation skill, but it needs review because it can store and send an API key and user media through configurable, insufficiently validated endpoints.

Review before installing. Use a dedicated low-privilege AI Hive API key, avoid custom `--base-url`, `AI_HIVE_BASE_URL`, or edited config hosts unless you fully trust them, and only upload media you are comfortable sending to the configured service. Prefer an isolated Python environment and a pinned dependency install.

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 (3)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/imagegen.py:90
Finding

Unrestricted API Endpoint Override Can Disclose the Bearer API Key

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/imagegen.py:279
Finding

Server-Controlled Upload and Download URLs Are Used Without Trust Validation

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Note
Location
SKILL.md:79
Finding

Runtime Dependency Installation Is Unpinned and Lacks Integrity Verification

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (8)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding

The skill invokes shell commands, installs a package, and runs a Python script with network, file, and environment access, but it does not declare any tool scope or permission boundaries. This creates unnecessary ambient authority: an agent may execute the skill with broader capabilities than intended, increasing the chance of unsafe file access, network egress, or command execution if the script or its inputs are abused.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest description and the entire skill guidance are written as mandatory Chinese-language instructions for generating JD e-commerce images, with no indication that users may choose another language or that the locale restriction is optional. This matches the policy category for language/locale constraints imposed without user opt-in.

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 advertised as a JD e-commerce image generation/editing tool, but the file also contains generic chat and video generation primitives plus broader OpenAPI access logic. This expands the skill’s effective capability beyond user expectations and least-privilege design, increasing the risk of unauthorized or unintended use if another component invokes these hidden functions or the file is repurposed.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The module docstring and all user-facing CLI guidance are written in Chinese, establishing a Chinese-only interaction pattern. The policy requires flagging language or locale constraints when the skill forces a specific language without opt-in or documented justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

This markdown file presents substantive skill documentation only in Chinese, which effectively forces a specific language on readers. Under the policy, language constraints should either be optional for the user or clearly justified as region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
96% confidence
Finding

The placeholder value for api_key includes Chinese text (请替换为你的APIKey), which imposes a specific language in user-facing configuration content. The file does not offer a language choice or explain that the skill is intended only for a Chinese-speaking or region-specific audience.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

User-info and wallet-balance retrieval are unrelated to the stated JD image-generation purpose and expose account metadata that is not necessary for core operation. Even if not directly exploitable on their own, unnecessary account-inspection features increase information exposure and violate capability minimization.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

Model listing and model lookup expose broader service discovery than required for a fixed-purpose JD image skill, allowing users or wrappers to enumerate capabilities beyond the advertised image workflow. In context, this is an overbroad feature set that increases attack surface and can aid misuse of the underlying platform.

Content

No source excerpt is available for this finding.