ClawHub
Back to skill

Security audit

Computer Usage Summary

Security checks for vulnerabilities and agentic risk

Overview

This skill locally summarizes ActivityWatch computer-usage data and discloses the main privacy-sensitive behavior.

Install only if you are comfortable letting the agent summarize local ActivityWatch records, which can reveal app usage and window-title context. Keep the API on 127.0.0.1, review any rules file before use, and save CSV/TSV reports only when you intend to create local files containing activity data.

SkillSpector

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 (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
84% confidence
Finding
The skill directs use of local file access, network access to a localhost API, and file output, but the metadata shown here does not declare corresponding permissions. Undeclared capabilities weaken reviewability and user consent because a caller may not realize the skill can read local data sources or write reports to disk.

Tp4

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding
The skill description promises privacy-preserving live-state and Screen Time fallbacks, but the analyzed behavior indicates those fallbacks are not actually implemented while additional capabilities such as local rules-based classification and writing output files are present. This mismatch can cause users to over-trust the skill's privacy model and consent to operations they did not expect, especially around local data handling and file creation.

Static analysis

No suspicious patterns detected.