ClawHub
Back to skill

Security audit

Ppt Light Corporate

Security checks for vulnerabilities and agentic risk

Overview

This is a local PowerPoint generation skill with disclosed local file writing and feedback memory, but no evidence of hidden collection, exfiltration, credential access, or destructive behavior.

Install if you are comfortable with a PPT helper that saves generated files locally and may retain feedback in markdown memory files. Avoid recording confidential slide content or business details in feedback memory, periodically review or delete those memory files, and verify any separately provided PPTX template/icon assets before use.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
87% confidence
Finding
The skill documentation indicates file read/write capabilities, including saving generated PPTs, backing up templates, and recording feedback to markdown files, but it declares no permissions. Undeclared filesystem access weakens least-privilege controls and can enable unauthorized reads or writes if the runtime trusts the metadata for policy enforcement.

Description-Behavior Mismatch

Medium
Confidence
90% confidence
Finding
The module includes a persistent 'learning loop' that writes feedback and best-practice data to disk, which exceeds the skill's stated scope of generating/improving PPTs. Even though the content written is not overtly sensitive by itself, this creates undeclared statefulness and local persistence, which can be abused to store user-derived content, leak information into future runs, or write into unintended filesystem locations if the caller controls memory_dir.

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
record_feedback() accepts an arbitrary memory_dir and appends data to a constructed path without restricting the destination to a safe base directory. In a skill context, this is dangerous because any caller able to influence memory_dir can cause local file writes outside the intended workspace, creating a path traversal/arbitrary file write primitive and persistent storage channel.

Context-Inappropriate Capability

Medium
Confidence
95% confidence
Finding
record_best_practice() has the same issue: it writes to a caller-influenced directory and persists structured content unrelated to core PPT rendering. This can be exploited to create or modify local files in unexpected locations and to retain user/project metadata across sessions without clear disclosure.

Static analysis

No suspicious patterns detected.