ClawHub
Back to skill

Security audit

Wiki Framework Builder

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent Chinese report-framework generator that reads a scoped JSON handoff and creates report outline deliverables without hidden network, credential, persistence, or destructive behavior.

Install only if a Chinese consulting/report-framework workflow fits your use case. Review generated Word and Markdown outputs before sending them externally, especially to confirm internal labels, assumptions, and confidence details were not accidentally included.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Tp4

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding
The skill promises to consume task-scope JSON, produce multiple deliverables, and prevent leakage of internal routing labels, but the documented behavior lacks concrete logic enforcing those guarantees. This can cause operators to trust outputs that may omit required artifacts or accidentally expose internal taxonomy and assumptions, creating integrity and information disclosure risks in client-facing deliverables.

Natural-Language Policy Violations

Medium
Confidence
96% confidence
Finding
The contract hard-codes Chinese section labels and requires them to be shown directly in chat without providing a language-negotiation path. In a multilingual or English-first workflow, this can cause the agent to ignore user language preferences, degrade usability, and create scope-confirmation errors when critical assumptions are presented in a language the user cannot reliably review.

Natural-Language Policy Violations

Low
Confidence
89% confidence
Finding
The JSON schema fixes framework depth values to Chinese enums, which introduces locale coupling into a machine-readable contract. This can break downstream integrations, create parsing ambiguity across systems, and force consumers to special-case one language even when the rest of the workflow is multilingual.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.