Security audit

ImgLink

Security checks for vulnerabilities and agentic risk

Overview

This is a simple documentation-only image URL skill, with the main caution that prompts and optional API keys are placed in shareable URLs.

Install only if you are comfortable sending prompt text to imglink.ai through URLs. Use anonymous access or a backend-held key for public/shared embeds, and avoid putting secrets, personal data, confidential business details, or reusable API keys in prompts or image URLs.

A.I.G

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

Warning

Location: SKILL.md:17
Finding: API Key Exposure Through Publicly Embedded Query Strings
Content: View full analysis
tags, Markdown, CSS, or anywhere an image URL is accepted. - Images are generated on first request, then cached permanently. Same prompt + version = same image. - URL-encode the prompt value. Use + for spaces. - Do not invent parameters that are not listed above. ## Examples ``` ``` ``` ### Technical Analysis The skill identifies `key` as an API key, places it in the URL query string, and recommends using the resulting authenticated URLs directly in HTML, Markdown, CSS, websites, presentations, and PDFs. Query-string credentials are not an appropriate mechanism for distributing long-lived secrets. The complete URL may be exposed through: - HTML or Markdown source viewed by recipients. - Published documents, presentations, and PDF source or metadata. - Browser history and bookmarks. - Web server, reverse proxy, CDN, monitoring, and analytics logs. - Collaboration systems that retrieve URLs to generate link previews. - Screenshots, support records, or copied image addresses. - Referrer information in environments whose referrer policy permits URL details to be propagated. Because every consumer must receive the complete image URL, embedding a non-anonymous key inherently distributes that credential beyond the key holder. The documented permanent caching behavior may also make prompt-derived resources persist beyond the intended use period, although the primary confirmed weakness is credential exposure. ## ...[truncated 1418 chars]
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The skill instructs users to place arbitrary prompts directly into a GET URL, which exposes prompt contents to the third-party service and potentially to browser history, logs, analytics systems, referrer headers, proxies, and shared documents. Because prompts may contain sensitive business data or personal information, the lack of an explicit privacy warning can lead users to disclose confidential content unintentionally.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.