Security audit

Sol on X

Security checks across malware telemetry and agentic risk

Overview

This markdown-only Solana/X content strategy skill has broad activation language, but it does not install code, access credentials, post automatically, or handle funds.

Install this only if you want help producing Solana/X growth content and public narrative research. Review generated posts before publishing, especially token commentary, reply bait, hot takes, and named-person references. The artifacts do not justify granting wallet, purchase, credential, or auto-posting access unless a separate tool or workflow explicitly asks and you intend that behavior.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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

High

Confidence: 93% confidence
Finding: The skill advertises extremely broad trigger phrases like 'tweet ideas,' 'growth on X,' 'Solana shilling,' and 'engagement farming,' which can cause the agent to invoke this skill for loosely related requests. Over-broad routing can override more appropriate skills, cause unexpected handling of user data, and increase the chance that high-risk social-engineering or manipulative content generation is used in contexts the user did not intend.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The skill explicitly instructs use of stored memory about the user's account positioning and goals without any user-facing notice or consent mechanism. In a profiling-heavy social growth context, silently reusing identity, audience, and business-positioning data can surprise users, expose sensitive preferences, and create privacy and trust risks if memory is stale, overbroad, or shared across contexts.

Natural-Language Policy Violations

Medium

Confidence: 88% confidence
Finding: The file sets a default writing style ('lowercase,' 'terse,' 'no marketing voice') that can steer outputs into a fixed persona without explicit user consent. In a social-growth skill this is especially relevant because it can override user-preferred tone, reduce transparency about stylistic manipulation, and nudge the model toward engagement-optimized behavior by default.

Natural-Language Policy Violations

Medium

Confidence: 91% confidence
Finding: The 'outsider tells' section instructs the model to avoid normal terminology and substitute insider jargon, effectively coercing a dialect shift. In this skill context, that makes outputs more manipulative because it helps impersonate an in-group voice and can obscure meaning for users who did not ask for culture-specific phrasing.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The file declares it should be used for /growth-strategy, /engagement-audit, and any positioning question, where any positioning question is an undefined catch-all trigger. Overbroad invocation criteria can cause the skill to activate outside its intended scope, increasing the chance of inappropriate use, prompt-routing mistakes, and accidental exposure of manipulative social-growth guidance in unrelated contexts.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The trigger phrase "what's hot" is overly broad and can match ordinary conversation unrelated to Solana narrative analysis, causing unintended skill invocation. In an agent setting, this can route user requests to the wrong skill, produce irrelevant web-search behavior, and increase the chance of prompt-crossing or unnecessary exposure to external content.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.