ClawHub
Back to skill

Security audit

Mission Control

Security checks across malware telemetry and agentic risk

Overview

This skill appears locally focused, but it uses very broad activation rules and can persist copied local file contents in a way users should review carefully.

Install only if you want a broad local planning gate that records workflow state. Avoid passing sensitive file paths as plan content, review the mission-control data directory periodically, and prefer explicit user confirmation before the skill reads files or activates on routine tasks.

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
93% confidence
Finding
The script treats the third argument as either raw content or a filesystem path and, if the path exists, reads that local file and copies its contents into the mission-control directory. In an agent skill context, this expands the tool from 'save provided plan text' into an arbitrary local file read primitive, which can expose sensitive files if untrusted input can influence the argument.

Vague Triggers

High
Confidence
95% confidence
Finding
The activation criteria are so broad that the skill can capture many normal tasks involving code, files, terminal use, or web research, creating an unnecessary mandatory workflow gate. In an agent environment, overbroad interception can be abused to redirect routine work into filesystem-heavy logging and approval flows, increasing prompt-control surface and causing denial of service or policy confusion.

Vague Triggers

Medium
Confidence
93% confidence
Finding
The automatic judgment logic uses ambiguous categories such as 'research web retrieval,' 'terminal operations,' and file extensions without clear boundaries, which invites inconsistent or overly aggressive triggering. In context, this makes the skill more dangerous because it is designed to insert itself before execution and can therefore reshape agent behavior based on fuzzy heuristics rather than explicit authorization.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.