Security audit

Decision Algorithm

Security checks across malware telemetry and agentic risk

Overview

The skill appears to be a legitimate decision-analysis helper, but its very broad triggers and investment-sizing outputs need review before installation.

Review this skill before installing. Use it only with explicit decision-analysis prompts, avoid relying on its investment-sizing output as financial advice, and restrict or confirm any web, shell, read, or write tool use unless you intentionally want the agent to perform research or local operations.

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 (6)

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The README encourages invocation through very broad, everyday phrases such as 'Should I...' and 'How risky is this opportunity?', which are likely to appear in normal user conversation. In agent environments that auto-route or auto-activate skills based on trigger phrases, this can cause the skill to activate unintentionally and influence sensitive life, financial, or relationship decisions without the user explicitly requesting this specific skill.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The analysis-depth examples use generic phrases like 'Help me quickly assess this decision' and 'Help me analyze this in detail', which are common requests unrelated to this specific skill. In systems that rely on phrase matching, these examples increase the chance of accidental triggering, causing unintended takeover of ordinary assistance flows and potentially steering high-stakes personal decisions through this framework.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The trigger phrases are very broad and overlap with ordinary conversation, which can cause the skill to activate in unintended contexts. Because the skill is allowed to use WebSearch, WebFetch, Bash, Read, and Write, accidental activation can lead to unnecessary tool use, unrequested external lookups, and expanded attack surface from benign user text.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The usage examples reinforce broad activation without clear boundaries, making it easy for the skill to engage on ambiguous or casual prompts. In this skill's context, that is more dangerous because the instructions explicitly say research 'must use tools' for fact-dependent questions, increasing the chance of unneeded browsing or command execution paths being entered.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The trigger list includes very broad everyday phrases such as "risk," "invest," and "should I," which can cause the skill to activate in many unrelated contexts. In a decision-making skill, unintended invocation is especially concerning because it may steer sensitive personal, financial, or relationship conversations into this framework without clear user intent, increasing the chance of overreach or inappropriate advice.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: This tool outputs explicit financial action guidance such as suggested investment amounts and position sizing without any disclaimer, suitability check, or warning that the results are only rough mathematical heuristics. In the context of a decision-making skill triggered by prompts like 'how much to invest' and 'risk,' users could reasonably rely on these outputs as actionable financial advice, increasing the chance of harmful real-world decisions.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal