ClawSentinel

Security checks across malware telemetry and agentic risk

Overview

ClawSentinel is a simple instruction-only skill for read-only security auditing, with no executable code or hidden persistence in the reviewed bundle.

Reasonable to install based on the provided artifacts. Treat its scan results as advisory, avoid relying on its marketing claims as proof of security coverage, and remember it may fetch public GitHub raw content when you explicitly ask it to audit a repository.

SkillSpector

By NVIDIA

Vulnerability Patterns

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
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Vague Triggers

Medium

Confidence: 83% confidence
Finding: The trigger phrase "audit this skill:" accepts pasted content without defining clear boundaries, size limits, or trusted input scope. In a security-auditing skill, this can enable prompt-injection or content-smuggling attacks where adversarial markdown instructs the model to ignore its guardrails, making the skill more dangerous because it is explicitly designed to process untrusted text.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal