ClawHub
Back to skill

Security audit

Blog Watcher

Security checks across malware telemetry and agentic risk

Overview

This blog/RSS helper does not look malicious, but it gives the agent broad command and file tool authority without explaining why that is needed.

Install only if you are comfortable reviewing each proposed command, file read, or file write before it runs. Prefer using it with explicit RSS/feed URLs and do not allow shell commands unless the agent explains a clear, task-specific reason.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

Context-Inappropriate Capability

Medium
Confidence
89% confidence
Finding
The skill declares unrestricted shell execution via bash even though its stated purpose is only to monitor blogs and RSS feeds. That unnecessary capability expands the attack surface: a prompt injection from fetched web content or an overly broad user request could cause arbitrary local command execution unrelated to feed monitoring.

Context-Inappropriate Capability

Low
Confidence
76% confidence
Finding
Broad web_search access exceeds the narrow function of monitoring specific blogs and RSS feeds, which should usually rely on explicit feed URLs or targeted fetching. While less severe than shell access, this capability can be abused to expand the skill's reach, follow malicious search results, or assist data exfiltration and prompt-injection chains.

Vague Triggers

Medium
Confidence
81% confidence
Finding
The trigger phrase "Help me with blog watcher" is broad and overlaps with ordinary help-style requests, increasing the chance the skill activates in contexts where the user did not specifically intend it. In a skill with unnecessary powerful tools like bash and web access, accidental or overly broad activation makes misuse and unsafe tool invocation more likely.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.