Image Fetch Toolkit

• Outcome: Pass
• Latest audit: May 27, 2026 · 10:55 PM UTC
• Version: 1.0.0

Overview

This is a broad but coherent image-search reference skill, with no hidden code or automatic actions, but users should be mindful of third-party API queries, credentials, licensing, and optional external tool installs.

Install this only if you are comfortable with an agent using third-party image, news, academic, and social/search APIs. Configure only the API keys you need, avoid sensitive or confidential search terms, review any optional external tools before installing them, and check licensing, attribution, and platform terms before reusing fetched images.

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (10)

Context-Inappropriate Capability

Medium

• Confidence: 89% confidence
• Finding: The skill’s stated purpose is image search and retrieval, but it also includes guidance for scraping e-commerce and social media sources, including scraper alternatives and platform-specific collection methods. That scope expansion increases the chance an agent will perform higher-risk collection against sites with sensitive content, ToS restrictions, or account-bound data, which is unnecessary for the core use case.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The manifest description says to use this skill for image retrieval 'for any purpose' and across many contexts, making activation criteria overly broad. Overbroad routing can cause the agent to invoke networked retrieval and third-party APIs in situations where a narrower local or privacy-preserving option would be safer.

Vague Triggers

Medium

• Confidence: 86% confidence
• Finding: The introductory text presents a comprehensive image retrieval toolkit without clear limits on source trust, legality, or user-consent requirements. In an agent setting, ambiguous scope increases the likelihood of misuse, unnecessary external calls, and selection of risky retrieval methods.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill strongly encourages sending search queries and URLs to numerous external providers but does not warn about privacy exposure, copyright/licensing limits, or third-party processing of user data. Users may unknowingly transmit sensitive prompts, internal URLs, or regulated content to external services.

External Transmission

Medium

• Category: Data Exfiltration

Content

```bash
# Search photos
curl -H "Authorization: Bearer $UNSPLASH_ACCESS_KEY" \
  "https://api.unsplash.com/search/photos?query=renewable+energy&per_page=10&orientation=landscape"

# Get a random photo

• Confidence: 77% confidence
• Finding: This example instructs the agent to send user-provided search terms to Unsplash, which is an external transmission of potentially sensitive user intent. In context this is core functionality, but it remains security-relevant because no guardrails are provided to prevent leaking private or confidential queries.

External Transmission

Medium

• Category: Data Exfiltration

Content

```bash
# Search photos
curl -H "Authorization: $PEXELS_API_KEY" \
  "https://api.pexels.com/v1/search?query=nature&per_page=10&orientation=landscape"

# Curated photos
curl -H "Authorization: $PEXELS_API_KEY" \

• Confidence: 76% confidence
• Finding: This example directs the agent to send search queries to Pexels, which exposes user intent to a third party. That is expected for an image search skill, but without privacy constraints it can leak sensitive requests or contextual information.

External Transmission

Medium

• Category: Data Exfiltration

Content

```bash
# Search photos (CC license only)
curl "https://api.flickr.com/services/rest/?method=flickr.photos.search&api_key=$FLICKR_API_KEY&text=sunset&license=1,2,4,5,9&per_page=10&format=json&nojsoncallback=1"

# Get photo URL: https://farm{farm}.staticflickr.com/{server}/{id}_{secret}.jpg
```

• Confidence: 74% confidence
• Finding: The Flickr search example performs an external query using user-supplied search terms and an API key. In a benign skill this is expected, but it still creates data-exposure risk when the skill lacks instructions to sanitize or avoid sensitive searches.

External Transmission

Medium

• Category: Data Exfiltration

Content

```bash
curl -H "Ocp-Apim-Subscription-Key: $BING_API_KEY" \
  "https://api.bing.microsoft.com/v7.0/images/search?q=cute+cats&count=10&imageType=Photo&size=Large"
```

### Tavily Extract (for webpage images)

• Confidence: 76% confidence
• Finding: This Bing image search example sends a query to a third-party provider and could disclose user interests or sensitive request details. The context makes it functional rather than malicious, but the absence of privacy warnings and minimization controls leaves a real exposure risk.

External Transmission

Medium

• Category: Data Exfiltration

Content

```bash
# Search papers
curl "https://api.semanticscholar.org/graph/v1/paper/search?query=protein+folding&limit=10&fields=title,url,openAccessPdf"
```

### Semantic Scholar API (Free)

• Confidence: 71% confidence
• Finding: The Semantic Scholar paper search example transmits research queries externally, potentially exposing confidential research interests or unpublished project topics. Because the skill targets academic use cases, that contextual risk is more meaningful than for generic stock-photo search.

External Transmission

Medium

• Category: Data Exfiltration

Content

### Semantic Scholar API (Free)
```bash
# Search papers with figure references
curl "https://api.semanticscholar.org/graph/v1/paper/search?query=transformer+architecture&limit=5&fields=title,url,openAccessPdf,figures"
```

### arXiv API (Free, no key needed)

• Confidence: 71% confidence
• Finding: This figure-related scholarly search also sends potentially sensitive research topics to an external service. In academic and enterprise settings, such queries can reveal strategic R&D directions or unreleased work.

Static analysis

No suspicious patterns detected.