Back to skill

Security audit

Check AI Crawler Access

Security checks for vulnerabilities and agentic risk

Overview

The verdict is benign with high confidence: the scanners report a low-severity privacy concern about sending a supplied site URL and optional page path to an external service, while the flagged behavior is considered related to the skill’s stated purpose.

Before installing, consider whether URLs you might submit could reveal private hostnames, sensitive page paths, or credentials. The scanner reports that these inputs may be sent to an external service without a privacy warning; the verdict is benign, but this finding is a reason to avoid submitting sensitive URLs.

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
Vulnerability Patterns
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

The workflow sends the supplied site URL and optional page path to a third-party service without a privacy warning.

Medium
Category
Output Handling
Confidence
91% confidence
Finding

The workflow instructs the agent to transmit a supplied URL and optional page path to an external service without checking for sensitive information or notifying the user of that disclosure. URLs can contain private hostnames, confidential document paths, or query-string credentials, which could consequently be exposed to the service. This is a legitimate dependency for the advertised functionality, and ordinary public-site checks pose little confidentiality risk, but the skill does not constrain inputs to that case.

Content

No source excerpt is available for this finding.