ClawHub
Back to skill

Security audit

SenseAudio-TTS

Security checks for vulnerabilities and agentic risk

Overview

This is a coherent SenseAudio cloud text-to-speech helper that uses an API key and sends text to SenseAudio as expected for its purpose.

Install only if you intend to use SenseAudio. Keep SENSEAUDIO_API_KEY scoped and private, avoid sending secrets, regulated data, or confidential text for speech generation unless authorized, and ensure any session or trace logging is access-controlled and excludes request text and credentials.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Context Leakage

High
Category
Data Exfiltration
Confidence
85% confidence
Finding

In a skill that operates with an API key and external service sessions, instructing the agent to log session and trace identifiers can expose correlatable metadata into logs, prompts, or downstream systems. Even if the values are not secrets themselves, they may enable cross-system tracking, aid incident reconstruction by an attacker with log access, or accidentally capture adjacent sensitive request context.

Content

Scanner excerpt · SKILL.md (reported line 44)May include surrounding context.

md
4. Harden for production:
- Handle auth and parameter errors.
- Add retry/backoff for transient network failures.
- Log session and trace identifiers.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger text is broad enough to activate on generic requests like TTS generation, voice selection, streaming speech, or troubleshooting, which could cause this skill to be invoked outside a narrowly scoped SenseAudio context. Over-broad routing increases the chance that credentials, network actions, or provider-specific instructions are applied when another tool or a safer generic path would be more appropriate.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

The skill references external HTTP and WebSocket endpoints, which means any text processed through the skill leaves the local environment and is transmitted to a remote service. In isolation this is expected for a cloud TTS integration, but within an agent ecosystem it creates a real data-exposure boundary that can leak sensitive user content if not clearly disclosed and controlled.

Content

Scanner excerpt · references/tts.md (reported line 16)May include surrounding context.

md
## Endpoint Summary

- HTTP/SSE endpoint: `POST https://api.senseaudio.cn/v1/t2a_v2`
- WebSocket endpoint: `wss://api.senseaudio.cn/ws/v1/t2a_v2`
- Auth: `Authorization: Bearer <API_KEY>`

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The document instructs users to send arbitrary text and authenticate to a third-party TTS provider but does not warn that input text may contain sensitive or regulated data. In an agent skill context, this omission can cause accidental exfiltration of prompts, personal data, or confidential business content to an external service.

Content

No source excerpt is available for this finding.