ClawHub
Back to skill

Security audit

Prove Before Act

Security checks across malware telemetry and agentic risk

Overview

This documentation-only skill describes a coherent proof-auditing integration that sends local hashes and selected metadata to an external certification service, with the main privacy and payment risks disclosed.

Install only if you want proof hashes and selected metadata anchored with a third-party service and potentially visible publicly forever. Use placeholder or non-sensitive filenames and author names, never send raw files or secrets, protect any pm_ API key, and enable x402 payments only with explicit spending limits and approval rules.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The documentation explicitly describes multiple public endpoints that expose certification metadata, filenames, proof documents, PDF certificates, and badges without any discussion of privacy implications or data minimization. In an accountability/certification product, these records can leak sensitive filenames, authorship, timestamps, and business activity patterns, which may enable unwanted disclosure, correlation, or reconnaissance even if the underlying file contents are not exposed.

Missing User Warnings

Low
Confidence
87% confidence
Finding
The MCP tool documentation instructs agents to send file-derived metadata (`file_hash`, `filename`) and optionally a `webhook_url` to an external service, but it does not warn about privacy, data sharing, or SSRF-style risks from callbacks. In an agent context, even hashes and filenames can be sensitive, and an unconstrained webhook parameter can cause unintended outbound interactions or disclosure if supplied from untrusted input.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.