Back to skill

Security audit

TS Images

Security checks for vulnerabilities and agentic risk

Overview

This image-generation skill is purpose-aligned, but it sends a bearer token and prompt/image request data to fully configurable URLs without validating the destination.

Install only if you trust the source of AIZNT_PROXY_URLS and TS_TOKEN. Before use, verify the configured URLs point to the intended HTTPS service, avoid sending sensitive prompts or source images unless needed, and prefer a version that restricts allowed hosts before attaching the bearer token.

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 (1)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/client.js:2
Finding

Bearer Credential and Request Data Can Be Sent to Unrestricted Configurable Endpoints

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (10)

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The skill description focuses on image generation features, but the metadata also depends on credentials and proxy URL configuration that are not clearly disclosed as operational behavior. This mismatch reduces transparency for reviewers and users, making it easier to hide sensitive data access or network routing behavior behind an apparently harmless media-generation skill.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding

The skill declares access to environment-backed credentials and proxy URL configuration, and its documented commands clearly imply outbound network use, but it does not define any explicit tool scope such as permissions or allowed-tools. That creates an authorization ambiguity where the runtime may grant broader env/network access than users or reviewers expect, increasing the chance of secret exposure or unintended external calls.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The manifest description and the entire markdown instructions are written in Chinese, with no indication that other languages are supported or that Chinese is an intentional opt-in. This can violate language/locale policy when a skill implicitly requires a specific language without user choice or documented justification.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This code reads a bearer token from TS_TOKEN and service endpoint data from AIZNT_PROXY_URLS, which are sensitive configuration values. Although there is an internal comment and error handling, there is no user-facing disclosure, confirmation, or visible logging that the skill accesses credentials/environment data.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
78% confidence
Finding

loadClient obtains a token that is then attached to outgoing requests through authHeaders, indicating credential use for remote calls. Accessing and using credentials is a safety-relevant action under this rule, and this file lacks any comment, warning, or user-facing notice about that behavior.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

This code posts user-supplied JSON bodies to remote image and content-generation endpoints via fetchJson, which can include user or system data. For code files, network transmission should have some visible disclosure such as a prompt, log, or explanatory comment/docstring unless the warning is documented elsewhere, and this file provides none beyond terse usage syntax.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Low
Category
Not specified by scanner
Confidence
76% confidence
Finding

The broader skill context describes '通用图生图(OpenAI 风格)与 Gemini generateContent 文生图', which implies both image-to-image and text-to-image support. In this manifest file, the name/tagline/description only mention generic images and Gemini generateContent, without reflecting the image-to-image scope claimed for the skill, creating a mild intent mismatch at the manifest-description level.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

The tagline is presented in Chinese while the description is in English, but the manifest does not explain any language or locale policy. This can create a language/locale constraint or inconsistent user experience without explicit opt-in or justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

Natural-language strings in the file, including the top-level comment and all thrown error messages, are written only in Chinese. This imposes a single language on users or operators without offering a language choice or documenting a justified locale constraint.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The script's natural-language interface, including usage text and error messages, is written only in Chinese. The policy calls for flagging language or locale constraints when a specific language is forced without user opt-in or a documented justification.

Content

No source excerpt is available for this finding.