Everyday Help Activator

• Outcome: Review
• Latest audit: May 29, 2026, 9:29 PM
• Version:
  1.0.0

Overview

This skill is not malware, but it asks the agent to keep ongoing local records about real people and their reactions without privacy limits.

Install only if you are comfortable with the agent keeping local notes about real-world interactions. Avoid names, unique descriptors, exact locations, schedules, contact details, or private conversation history; prefer anonymous aggregate notes such as which request phrasing worked, and delete logs regularly.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Context-Inappropriate Capability

Medium

• Confidence: 94% confidence
• Finding: The skill explicitly instructs creation of per-person relationship files keyed by name or type and to update them over time. That exceeds the stated purpose of giving momentary scripts for small favors and creates unnecessary profiling of identifiable third parties, which can expose private behavioral data and normalize covert relationship tracking.

Description-Behavior Mismatch

Low

• Confidence: 88% confidence
• Finding: The skill is presented as a lightweight real-world scripting aid, but the instructions expand into ongoing recordkeeping and relationship management. This mismatch increases the chance that users or agents will collect and retain more third-party data than is reasonably necessary for the advertised function.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill tells the agent to log successful activations with details about person type, ask, outcome, script used, reaction, and reciprocity without any privacy notice or minimization guidance. In context, this creates an unnecessary dossier of third-party interactions from routine real-world encounters, increasing privacy and misuse risk.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: Creating `relationship-[name-or-type].md` files encourages persistent storage of identifiable information about neighbors, cashiers, and other real people with no consent, warning, or retention limit. Because the skill concerns influencing recurring social interactions, the context makes this more concerning: it supports manipulation and tracking of specific individuals over time.

Ssd 3

Medium

• Confidence: 93% confidence
• Finding: The instructions create a natural-language data trail about third parties, including their role, responses, and prior interactions. Even if intended for benign social effectiveness, collecting this information about strangers and regular contacts can be repurposed for profiling, targeted manipulation, or unauthorized sharing if the filesystem is accessed by others.

VirusTotal

65/65 vendors flagged this skill as clean.