MeetingClaw - Intelligent Meeting Minutes Assistant

• Outcome: Review
• Latest audit: May 29, 2026, 9:26 PM
• Version:
  1.0.0

Overview

This meeting-minutes skill appears purpose-built, but it sends sensitive meeting content to cloud services and can automatically delete saved workspace files, so it needs user review before installation.

Install only if the meetings you process are allowed to be sent to Volcengine and the configured OpenClaw/model provider. Use limited-purpose API keys, avoid highly confidential or regulated recordings unless your policies permit these processors, and keep backups of generated minutes because the cleanup code can remove old notes as well as audio.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 92% confidence
• Finding: The skill describes capabilities that access environment secrets, read/write local files, and send data over the network, but it does not declare corresponding permissions. That mismatch weakens user and platform transparency, making it easier for a skill handling sensitive meeting recordings and transcripts to access secrets or persist data without clear consent boundaries.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill processes potentially sensitive meeting audio and transcription content through external cloud services and stores both recordings and generated notes locally by default, but the user-facing description does not clearly warn about these data flows. In a meeting-minutes tool, this is especially risky because recordings may contain confidential business, legal, HR, or personal information, so lack of explicit disclosure can lead to privacy violations and unsafe use.

Missing User Warnings

Medium

• Confidence: 83% confidence
• Finding: The guide tells users to create and use a pre-signed TOS URL for uploaded audio, but does not explain that the URL is effectively a bearer token granting temporary access to potentially sensitive meeting recordings. In this skill's context, the audio likely contains confidential business discussions or personal data, so accidental logging, sharing, or over-broad URL lifetime could expose sensitive content to unauthorized parties.

Ssd 3

Medium

• Confidence: 96% confidence
• Finding: Untrusted transcript text is inserted verbatim into the LLM prompt and the resulting minutes are then saved to disk and printed to stdout. Because transcripts may contain sensitive spoken content, secrets, or prompt-injection-style text spoken aloud or embedded in the transcription, the model may reproduce or overexpose confidential information in generated minutes, logs, or terminal history; this is especially risky because meeting data is inherently sensitive and the skill persists both raw audio and generated summaries.

VirusTotal

66/66 vendors flagged this skill as clean.