clawdio

• Outcome: Review
• Latest audit: May 28, 2026 at 9:11 AM UTC
• Version:
  1.1.0

Overview

This documentation-only skill is transparent about its paid report API, but it encourages automatic wallet-funded purchases without clear per-purchase approval or spending limits.

Install only if you are comfortable connecting a wallet-capable agent to this paid API. Use a dedicated low-balance wallet, require explicit approval before every purchase, browse the free catalog first, verify the report ID and $1.49 USDC price, and avoid sharing or storing speaker-attributed transcripts unless you have a clear basis to do so.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• 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)

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill advertises analysis of Twitter Spaces and voice conversations and explicitly returns participant lists, speaker-attributed quotes, timestamps, and transcripts, but the description does not clearly warn users that conversation-derived content and participant identifiers are sent to and processed by an external paid service. This creates a meaningful privacy and consent risk because an agent operator may invoke the skill on sensitive audio-related material without understanding the disclosure of personal or attributed data to a third party.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The documentation normalizes a charge-triggering GET request and says an x402 wallet will handle payment automatically, but it does not prominently warn that simply following the URL may initiate a paid transaction. In an agent-oriented API, this is dangerous because automated tools, crawlers, or LLM agents may treat GET as safe/read-only and trigger unintended purchases.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The purchased artifacts include participant identities, Twitter handles, avatars, attributed quotes, and full transcript content, but the API reference does not include any privacy, consent, retention, or downstream-use warnings. Because this skill is explicitly designed for market intelligence extraction from voice conversations, the context increases sensitivity: it encourages redistribution and analysis of personally attributable speech data that may create privacy, compliance, and reputational risks.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The integration guide explicitly states that payment and purchase behavior is 'fully automatic' and that an agent can 'browse and purchase' reports automatically, but it does not include any warning, approval requirement, spending limit guidance, or confirmation UX. In an agent setting, this can cause unintended on-chain spending if the agent is pointed at the service or follows links/tool outputs autonomously, especially because the payment flow is designed to be seamless via wallet-backed request signing.

VirusTotal

64/64 vendors flagged this skill as clean.