ClawHub
Back to skill

Security audit

AI播客生成-免费版

Security checks across malware telemetry and agentic risk

Overview

This podcast skill is mostly coherent, but its trigger is too broad for a tool that sends user text and PDF URLs to an external service.

Install only if you are comfortable sending the source text or PDF URL to MagicPodcast. Use it for non-sensitive podcast generation, avoid confidential or regulated documents, and be careful because its trigger language may match broader document-processing tasks than intended.

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)

Vague Triggers

Medium
Confidence
83% confidence
Finding
The trigger condition is overly broad for a skill that sends user-provided text or URLs to an external podcast API. Because it can activate on generic file processing, document conversion, or content extraction requests, users may invoke it in contexts involving sensitive documents without realizing their data will be transmitted off-platform. The skill context increases risk because the advertised use cases overlap with common enterprise and personal document workflows.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill’s main flow does not prominently warn that pasted text and PDF URLs are sent to a third-party API. Users may provide confidential notes, reports, or internal links under the assumption that processing is local, leading to unintended external disclosure. This is more dangerous in this skill because its core purpose is document transformation, which commonly involves sensitive content.

External Transmission

Medium
Category
Data Exfiltration
Content
# 从文本创建播客
payload="$(jq -n --arg text "$SOURCE_TEXT" --arg language "$LANGUAGE" '{text:$text,language:$language}')"

curl -sS -X POST "$MAGICPODCAST_API_URL/agent/v1/podcasts/text" \
  -H "Content-Type: application/json" \
  -H "x-api-key: $MAGICPODCAST_API_KEY" \
  --data-binary "$payload"
Confidence
95% confidence
Finding
This command transmits user-supplied source text to an external service using an API key. Even though the shell construction uses jq safely and does not show obvious command injection, the security issue is data exfiltration risk: sensitive user content can leave the local environment and be processed by a third party. In the context of a broadly triggered document-processing skill, this external transmission is materially risky.

External Transmission

Medium
Category
Data Exfiltration
Content
payload="$(jq -n --arg pdfUrl "$PDF_URL" --arg language "$LANGUAGE" '{pdfUrl:$pdfUrl,language:$language}')"

curl -sS -X POST "$MAGICPODCAST_API_URL/agent/v1/podcasts/pdf" \
  -H "Content-Type: application/json" \
  -H "x-api-key: $MAGICPODCAST_API_KEY" \
  --data-binary "$payload"
Confidence
94% confidence
Finding
This command sends a user-provided PDF URL and language setting to an external API for remote processing. Although the URL is lightly validated for scheme format, the key concern is third-party disclosure and possible transmission of internal or sensitive document locations. The danger is heightened because users may treat this as a generic document tool and submit links they would not knowingly share externally.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.