ClawHub
Back to skill

Security audit

pptgenerator

Security checks for vulnerabilities and agentic risk

Overview

The skill appears to be a legitimate PPT generation/editing integration, but it can automatically reuse and send prior presentation history to an external service without clear user opt-in.

Install only if you are comfortable with PPT content and related metadata being sent to pptultra.com. Avoid using it with confidential presentations until implicit history reuse is disabled or made opt-in, and prefer passing the exact current PPT file explicitly for edit operations.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding
The skill declares no permissions, but its instructions clearly require network access (`requests`), reading a local PPT JSON file via `--current-ppt`, and likely writing generated output. This creates a transparency and policy-enforcement gap: a host may allow the skill to run without surfacing to users that local files and outbound network access are involved.

Ssd 3

Medium
Confidence
93% confidence
Finding
For edit flows, the code automatically loads the most recent history record and sends it as `current_ppt` context when the user does not explicitly provide one. This can unintentionally leak prior users' prompts, share URLs, summaries, and PPT metadata into a later request to the external service, causing cross-session data exposure. The skill context makes this more dangerous because presentation content often contains business, educational, or personal data.

Static analysis

No suspicious patterns detected.