Security audit

Eric's Second Brain

Security checks across malware telemetry and agentic risk

Overview

This is a broad advice-routing knowledge skill; its main risk is over-activation, not hidden code or data access.

Install only if you want an opinionated, Eric-branded framework to route broad engineering, career/life, and AI/innovation questions. Be aware it may activate on generic topics and shape answers through its personal frameworks; it does not appear to add executable code, credential access, or background behavior.

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 (9)

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The README frames this skill as a universal first-entry point for engineering, life, cognition, innovation, and AI questions, but it does not define activation boundaries, exclusions, or precedence rules. In agent environments, overly broad invocation criteria can cause the skill to trigger on inappropriate prompts, route sensitive topics incorrectly, or overshadow more specialized skills, increasing the chance of unsafe or low-quality guidance.

Vague Triggers

High

Confidence: 95% confidence
Finding: The skill description claims it should be the 'first entry point' for essentially any question involving engineering, life, cognition, innovation, or AI, which is broad enough to match a large share of normal user requests. This can cause unintended invocation, overshadow more specific skills, and route sensitive or unrelated requests through a persona-driven meta-skill that injects subjective guidance and domain-spanning instructions without clear boundaries.

Vague Triggers

High

Confidence: 97% confidence
Finding: The routing rule explicitly states that 'any problem' should enter this skill first, creating an all-encompassing trigger that can capture broad categories of user requests before more appropriate, narrowly scoped skills are selected. In context, this is more dangerous because the skill is designed as a meta-router and authority layer, so overbroad activation can systematically hijack orchestration and propagate its own framing, templates, and subjective heuristics into many interactions.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The manifest defines very broad trigger phrases such as general engineering, life choice, career, AI, innovation, and cross-domain analysis terms. This can cause the skill to activate in many unrelated conversations, increasing the chance of unintended routing, context capture, and inappropriate influence over user interactions beyond the user's clear intent.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger examples are broad, natural-language phrases that overlap with ordinary user requests, so the meta-skill may capture inputs that should instead go to narrower or safer specialist skills. Because this skill is explicitly the 'first entry' and performs cross-domain routing, misrouting can systematically shape downstream behavior, causing irrelevant capability loading, prompt confusion, or unintended access to broader instruction sets.

Vague Triggers

Medium

Confidence: 96% confidence
Finding: The keyword routing table relies on single broad feature words such as '人生', '创新', and '技能' without disambiguation, making accidental or adversarial triggering easy. In a central orchestration skill, this increases the chance of over-broad skill activation and cross-domain prompt injection effects, where benign user text containing generic terms routes into powerful or unrelated skill paths.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The trigger phrases in this section are very broad everyday terms such as life, wealth, workplace, human nature, and history, which can easily match unrelated user requests and cause unintended routing into powerful meta-skills. In a top-level entry skill, overbroad activation increases the chance of scope creep, misleading responses, and accidental invocation of downstream capabilities without clear user intent.

Vague Triggers

Medium

Confidence: 83% confidence
Finding: The ecosystem map defines recognition triggers and cross-domain routing but does not specify hard activation boundaries, tie-break rules, or safety constraints for when multiple skills match. In a meta-router skill advertised as the first entry point for many topics, this ambiguity can lead to unintended activation chains, overreach into unrelated domains, and unpredictable behavior.

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger keyword table is broad enough to match many ordinary engineering and AI-related prompts, which can cause the skill to activate as a default entry point for a wide range of conversations. In this skill’s context, that increases the chance of unintended routing, prompt hijacking of unrelated tasks, and over-collection or over-application of this skill’s instruction set beyond the user’s actual intent.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.