Back to skill

Security audit

Ntriq X402 Pii Detect Batch

Security checks for vulnerabilities and agentic risk

Overview

This skill is a remote paid PII-detection API wrapper, but it under-discloses that raw sensitive text is sent to a third-party service before masking.

Review this carefully before installing or using it with real personal, customer, employee, financial, health, or regulated data. Only submit text you are authorized to send to this provider, confirm the x402 payment details before attaching a payment header, and look for clear privacy, retention, deletion, logging, and data-processing terms from the service operator.

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
SKILL.md:15
Finding

Unverified Third-Party Transmission of Sensitive Personal Data

Content
View full analysis
{ "texts": [ "Contact John Smith at john@example.com or 555-0123", "Invoice sent to 123 Main St, Springfield" ], "mask": true } ``` ``` ### Technical Analysis The documented workflow instructs the agent or user to send text containing names, email addresses, telephone numbers, Social Security numbers, addresses, and potentially other regulated information to the externally operated `x402.ntriq.co.kr` service. The Skill does not provide safeguards requiring informed user authorization before disclosure. It also does not document the service operator's data-retention, deletion, secondary-use, logging, access-control, breach-response, or jurisdictional practices. Consequently, users cannot determine how submitted PII will be handled after it leaves their environment. The statement that inference is “100% local” is ambiguous and potentially misleading. The documented HTTP request demonstrates that processing is not local to the user or agent environment. “Local” apparently refers to infrastructure controlled by the remote provider, such as its Mac Mini, rather than client-side processing. This wording may cause users to transmit sensitive information under an incorrect privacy assumption. The request additionally carries an `X-PAYMENT` authorization header. Although no direct theft mechanism is present in the reviewed file, this combines sensitive-data transmission and payment authorizat ...[truncated 1765 chars]
Remediation
View remediation
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
97% confidence
Finding

This skill explicitly asks users to send up to 500 raw text inputs containing highly sensitive PII to a remote third-party service, but it does not warn users that the data leaves their environment or describe handling, retention, or privacy safeguards. In the context of a PII-detection tool, that omission is particularly dangerous because users may assume the operation is privacy-preserving while actually transmitting regulated personal data over the network to an external processor.

Content

No source excerpt is available for this finding.