ClawHub
Back to skill

Security audit

调用senseaudio asr的课堂转译助手，将英文课堂录音、讲座视频、组会音频等内容自动转写为文本，生成中文总结

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it claims: it turns user-selected class audio or video into notes using disclosed cloud services and optional exports.

Install only if you are comfortable uploading selected recordings to SenseAudio and, when enabled, sending notes or transcripts to a trusted OpenAI-compatible endpoint or Notion. Use scoped tokens, avoid confidential or unauthorized recordings, and do not set OPENAI_BASE_URL to an endpoint you do not fully trust.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Taint TrackingDirect Taint Flow, Variable-Mediated Taint Flow, Credential Exfiltration Chain
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Tainted flow: 'base_url' from os.getenv (line 401, credential/environment) → requests.post (network output)

Critical
Category
Data Flow
Content
if not api_key:
        raise RuntimeError("Missing OPENAI_API_KEY for openai-compatible summarization")

    resp = requests.post(
        base_url.rstrip("/") + "/chat/completions",
        headers={"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"},
        json={"model": model, "messages": messages, "temperature": 0.2},
Confidence
98% confidence
Finding
`OPENAI_BASE_URL` is taken directly from the environment and used as the destination for a request that includes the bearer token and full transcript-derived content. If that environment variable is changed to an attacker-controlled host, the skill will exfiltrate sensitive transcripts and the API key to that host, creating a clear SSRF/credential-leak path.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill describes sending classroom audio, derived media from webpages, and possibly transcript content to external services, but it does not prominently warn users about privacy implications or data disclosure. Classroom recordings can contain sensitive personal, academic, or confidential information, so silent transmission to third parties can violate user expectations and policy requirements.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill uploads user-provided classroom audio to an external ASR provider without any in-code notice, consent gate, or data-handling disclosure. Because recordings may contain personal data, student voices, or protected educational content, silent third-party transmission creates a real privacy and compliance risk.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
When `openai-compatible` summarization is selected, the code transmits transcript content to another external service without any explicit warning or confirmation. Transcripts from lectures or classrooms can contain sensitive personal, academic, or proprietary information, so undisclosed sharing meaningfully increases privacy risk.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
Exporting generated notes to Notion sends user-derived content to a third-party platform without any explicit warning in the code path. Even though export is optional, there is no disclosure of what content is uploaded or its sensitivity, which is risky for educational recordings and summaries.

Static analysis

No suspicious patterns detected.