ClawHub
Back to skill

Security audit

Audio Stream Upload

Security checks across malware telemetry and agentic risk

Overview

This skill is an audio upload helper whose network upload behavior is expected, but users should treat the reliability claims and credential handling guidance cautiously.

Install only if you intend to upload the chosen audio files and metadata to the documented third-party streaming API. Use environment variables or a secret manager for keys, avoid pasting real secrets into transcripts or shell history, verify public/private metadata settings, and do not rely on the advertised resumable upload or automatic backoff without testing the code first.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill advertises capabilities such as resumable uploads and automatic rate-limit detection/backoff, but the provided implementation does not actually implement those protections. This is dangerous because operators may rely on reliability and safety guarantees that do not exist, leading to repeated retransmission of sensitive media, failed jobs, or unsafe retry behavior against an external service.

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The function claims to support breakpoint resume, but the code only references an undefined progress method and does not show any durable checkpointing or verified resume logic. Users may trust it for large or sensitive uploads, but interruptions could cause duplicate or partial transmission to the external API without reliable recovery.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill is designed to upload local audio files and metadata to a third-party API, but it does not prominently warn that local content will leave the host system. This can cause unintended disclosure of proprietary media, internal training content, or embedded metadata when an agent executes the workflow on a user's behalf.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The examples instruct users to place public and secret keys in HTTP headers and shell commands without strong guidance on secure handling, redaction, or avoiding logs/history exposure. In agent or terminal environments, this can lead to credential leakage through process listings, shell history, transcripts, or debug output.

External Transmission

Medium
Category
Data Exfiltration
Content
if config:
                create_data.update(config)
            
            resp = requests.post(
                f'{self.base_url}/videos/create',
                headers={**self.headers, 'Content-Type': 'application/json'},
                json=create_data
Confidence
94% confidence
Finding
requests.post( f'{self.base_url}/videos/create', headers={**self.headers, 'Content-Type': 'application/json'}, json=

External Transmission

Medium
Category
Data Exfiltration
Content
file_size = os.path.getsize(file_path)
    
    # 创建音频对象
    resp = requests.post(f'{self.base_url}/videos/create',
        headers={**self.headers, 'Content-Type': 'application/json'},
        json={'title': title, 'type': 'audio'})
    audio_id = resp.json()['data']['id']
Confidence
94% confidence
Finding
requests.post(f'{self.base_url}/videos/create', headers={**self.headers, 'Content-Type': 'application/json'}, json=

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.