Back to skill

Security audit

网盘资源聚合搜索(百度/夸克/阿里云盘)

Security checks for vulnerabilities and agentic risk

Overview

The skill is marked suspicious because scanners report that it sends search terms to external services while bypassing configured proxies, which may expose queries and bypass monitoring.

Before installing, consider whether search terms may be sensitive and whether bypassing your configured proxy is acceptable, since the scanner says queries go to third parties without a privacy warning. Review the unresolved references and uninspected image as gaps in the evidence; the scans do not establish that the skill is safe.

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 (3)

The workflow sends search terms to third-party services and mandates bypassing the system proxy without warning about the privacy implications.

Medium
Category
Output Handling
Confidence
95% confidence
Finding

The skill explicitly requires ProxyHandler({}), bypassing configured proxies when sending search terms to external services. External queries are expected for this search skill, but mandatory direct connections can circumvent proxy-based monitoring or filtering and expose the client's direct egress IP; sensitive queries also reach third-party operators without a privacy warning. The stated troubleshooting purpose suggests negligence rather than malicious intent, and HTTPS still protects request contents in transit.

Content

No source excerpt is available for this finding.

Docker image referenced without tag or digest: 'docker run -d'.

Medium
Category
MCP Rug Pull
Confidence
75% confidence
Finding

Docker image references without a specific tag (:latest is implicit) or digest (@sha256:...) can be silently replaced by a malicious image.

Content

No source excerpt is available for this finding.

The skill requires Chinese-language error messages without offering a language choice.

Low
Category
Output Handling
Confidence
99% confidence
Finding

Line 250 explicitly requires an actionable Chinese prompt for each error condition. This imposes a specific output language without user opt-in or a documented locale-specific justification; it is not merely a request for clear wording.

Content

No source excerpt is available for this finding.