ClawHub
Back to skill

Security audit

Project Ghost

Security checks for vulnerabilities and agentic risk

Overview

This skill is a simple external web-reading API integration, with privacy caveats but no hidden local code, persistence, or deceptive behavior in the artifact.

Install only if you are comfortable sending requested public URLs and resulting page-processing data to Project Ghost's hosted API. Do not use it for private intranet pages, authenticated content, URLs containing secrets or tracking tokens, or sensitive legal/business material unless you have reviewed the provider's terms and data handling practices.

A.I.G

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The usage examples are broad natural-language phrases like 'Read this page and tell me what it's about' and 'What are the top stories on Hacker News right now?', which could be triggered during ordinary conversation without the user realizing an external skill is being invoked. In this skill's context, that matters because invocation can cause third-party network access and transmission of user-supplied URLs/content targets to an external service.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill tells users to set an API key and call a hosted endpoint, but it does not clearly warn that requested URLs and derived page content are sent to an external API provider for processing. This omission creates a transparency and privacy risk, especially if users provide sensitive, internal, or tracking-bearing URLs assuming the analysis happens locally.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
91% confidence
Finding

The skill explicitly sends data to an external endpoint using a bearer API key and a user-provided URL. External transmission is expected for this product, but it is still a real security concern because it can expose browsing targets, embedded query secrets in URLs, and extracted content to a third-party service if users are not adequately warned or restricted.

Content

Scanner excerpt · SKILL.md (reported line 50)May include surrounding context.

Set your GHOST_API_KEY environment variable, then call:

bash
curl -X POST https://project-ghost-production.up.railway.app/distill \
  -H "Authorization: Bearer $GHOST_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://apple.com"}'

Static analysis

No suspicious patterns detected.