Back to skill

Security audit

浙江企业工商信息查询(按名核验)

Security checks for vulnerabilities and agentic risk

Overview

This skill coherently looks up Zhejiang enterprise registry data through a disclosed remote service, with a local credentials file used for product access.

Install this if you are comfortable sending enterprise names or credit codes to the listed third-party service and storing this product's API/session data in ~/.odf/zj-enterprise-registry.json. Review or delete that file if you no longer want the saved access state retained.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (5)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger terms are broad and include common enterprise-information phrases, which raises the chance the skill will be invoked for generic business queries outside the user's specific intent. That can cause unnecessary network access, disclosure of third-party business data, or routing users into a specialized tool when a normal answer would suffice.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The script contacts external endpoints under sdk.tokendidi.cn, including a delivery API and purchase URL. Because no manifest or purpose is available, this network access is an unjustified capability relative to the declared intent baseline, which is effectively absent.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The script stores API keys and session cookies in a JSON file under the user's home directory and later reuses them in requests to a third-party service. Persisting bearer-style credentials and cookies locally without protection or clear user consent increases the risk of credential theft, session hijacking, and unintended long-term tracking if the file is read by other local processes or users.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The code writes credential material to a file in the user's home directory without any user-facing warning, consent flow, or visible permission hardening. Even if the intent is convenience, silent plaintext storage of API keys and cookies can expose secrets to other local software, backups, shared accounts, or accidental disclosure.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

Natural-language fields and instructions are predominantly Chinese, and the usage guidance assumes Chinese-language interaction. The file does not explicitly state that the skill is intentionally limited to Chinese users or offer an alternative language/locale choice beyond a brief English description field.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.