ClawHub
Back to skill

Security audit

ppt-style-matcher

Security checks across malware telemetry and agentic risk

Overview

The skill is mostly a disclosed PowerPoint style-matching tool, but it includes an overbroad whole-deck text replacement step that could change unrelated business content.

Review before installing if you work on important or client-facing decks. Use it only on copies of PPT files, specify the exact pages to modify, and do not allow the global residual keyword replacement unless you have reviewed the replacement list and want whole-deck changes.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding
The skill clearly directs reading and parsing local PPT files via python-pptx and analysis scripts, yet no permissions are declared. That creates a trust and enforcement gap: an agent may access user files without an explicit capability declaration or user-aware approval boundary.

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The documented residual-cleanup step performs global text replacement across all slides, which goes beyond style matching and page generation. This can silently alter unrelated content, names, or branding across the entire deck, causing integrity loss and potentially damaging business documents.

Description-Behavior Mismatch

Medium
Confidence
83% confidence
Finding
The workflow includes 'search and synthesize new content' even though the skill is presented as a style-matching and slide-generation tool. This expands behavior from formatting into autonomous content generation, which can introduce untrusted external information, hallucinated facts, or data-handling paths the user did not expect.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.