Security audit

Social Posting

Security checks across malware telemetry and agentic risk

Overview

This skill is not malicious, but it can publish to real social accounts and may fall back to shared global credentials without clear identity confirmation safeguards.

Install only if the host app clearly shows the final content, platforms, schedule, and exact posting identity before every publish or schedule action. Avoid shared global provider keys for user-initiated posts unless they are deliberately restricted to a known system account with explicit approval and tenant isolation.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The trigger phrases are broad enough to fire on generic requests like 'social post' or 'publish to platforms', which can cause unintended invocation of a capability that performs real external actions. In a skill that can publish content to multiple third-party accounts, accidental activation materially increases the risk of unauthorized or surprising posting behavior.

Missing User Warnings

Medium

Confidence: 97% confidence
Finding: The documented failover order allows the service to fall back from user credentials to global environment-variable credentials without an explicit user warning or consent checkpoint. In this context, that is dangerous because a user request could trigger posting through shared application-owned accounts, resulting in cross-tenant posting, brand-account misuse, or publication under the wrong identity.

VirusTotal

No VirusTotal findings

View on VirusTotal