Back to skill

Security audit

语音转文字 ElevenLabs STT

Security checks for vulnerabilities and agentic risk

Overview

The verdict is benign with medium confidence: the scanners found no AIG issues, but Skillspector flagged a mismatch between the stated transcription purpose and the documented dLazy CLI login flow, and its coverage was partial.

Consider whether you trust the dLazy CLI and its login and data-handling practices before installing. The scan did not inspect its implementation, and Skillspector's analysis was partial because several local references were missing; the benign verdict is not a guarantee that the skill or its dependencies are risk-free.

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
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • 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 (5)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied code only runs dlazy login. Nothing in the chunk connects that command to ElevenLabs or shows it supporting transcription, language detection, or diarization. Its visible behavior therefore materially differs from the declared purpose. The implementation and effects of the external command are not provided.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied code contains only dlazy login; it shows no ElevenLabs integration, transcription, language detection, or diarization. Authentication can be a supporting implementation detail, but the supplied chunk establishes no connection between this login command and the declared transcription service. The command's precise effects depend on the installed dlazy implementation.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
87% confidence
Finding

The command broadly matches the transcription and diarization purpose, and its execution and saving options are supporting functionality rather than unrelated capabilities. However, its documented language behavior conflicts with the advertised automatic detection. This assessment is limited to the supplied help text; the underlying implementation and use of scribe_v1 cannot be verified.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
80% confidence
Finding

The supplied chunk is a generic CLI invocation rather than an identifiable speech-to-text workflow. It shows no ElevenLabs integration, audio input, language detection, or diarization. An external CLI could support those features, but neither the unspecified command nor the package's implementation establishes that connection here. The finding is limited to the supplied chunk; the package's downstream behavior is unknown.

Content

No source excerpt is available for this finding.

Whitespace Padding

Medium
Category
Prompt Injection
Confidence
70% confidence
Finding

Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.

Content

Scanner excerpt · SKILL-cn.md (reported line 122)May include surrounding context.

md
## 错误处理

| Code | 错误类型                         | 示例信息                                                                                                                 |
| ---- | -------------------------------- | ------------------------------------------------------------------------------------------------------------------------ |
| 401  | 未授权 (API Key缺失或无效)       | `ok: false, code: "unauthorized", message: "API key is missing or invalid"`                                                              |
| 501  | 缺少必填参数                     | `error: required option '--prompt <prompt>' not specified`                                                               |