Security audit

Taku Reflect

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed local reflection tool that can save project learnings and retrospectives, with no evidence of hidden exfiltration or destructive behavior.

Install this if you want project-local reflection memory. Review any proposed entries before saving, avoid storing secrets or sensitive personal data in learnings, and only approve AGENTS.md or CLAUDE.md bootstrap changes if you want future agents to consult those local learnings.

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 (2)

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The trigger condition includes vague emotional cues like user satisfaction or frustration after work completion. Broad triggers can cause unintended invocation, which is risky here because the skill can search local state, write learnings, run git analysis, and potentially propose or perform persistent project-file updates after consent flows.

Vague Triggers

Medium

Confidence: 87% confidence
Finding: The trigger description is overly broad and includes common phrases like creating, editing, testing, or verifying a skill, which can cause the skill to activate in situations beyond its intended scope. In an agent system, ambiguous invocation increases the chance of unintended self-referential behavior, incorrect tool routing, or applying documentation-generation logic to security-sensitive review tasks.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.