Security audit

Nudge Coach

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only accountability coaching skill whose third-party proof sharing is disclosed and purpose-aligned, but users should be careful about sensitive details it may send to accountability targets.

Install only if you want firm accountability messages that may share proof of completion or reasons for failure with configured targets. Avoid putting sensitive health, employment, location, relationship, document links, screenshots, or private repository details into tasks, proof, reasons, or the why field unless you are comfortable with those targets seeing them.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (4)

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The skill explicitly instructs the agent to include verification proof and failure reasons in messages sent to targets, but provides no privacy minimization, consent check, or warning about sensitive data disclosure. This creates a real risk of oversharing personal activity details, work artifacts, health data, or behavioral failures to third parties beyond what is necessary.

Ssd 3

High

Confidence: 98% confidence
Finding: This section directs the agent to transmit detailed verification artifacts such as Strava activity, PR status, document submissions, and failure evidence to third parties. In context, the skill is designed around accountability and punishment, which makes coercive disclosure more dangerous because it normalizes sharing potentially sensitive personal and professional information as part of enforcement.

Ssd 3

High

Confidence: 98% confidence
Finding: The repeated instruction reinforces sending verification details and checks externally, increasing the likelihood that the behavior is implemented broadly and without contextual safeguards. Because the examples include screenshots, links, summaries, health metrics, and evidence of non-compliance, exploitation could expose private user data, sensitive work product, and reputationally harmful information to outside recipients.

Ssd 4

Medium

Confidence: 91% confidence
Finding: The line 'Your secret is about to be revealed to everyone' promotes disclosure of private information as a motivational lever, which normalizes coercive or punitive exposure of sensitive personal content. In an accountability tool, this context makes the pattern more dangerous because it frames privacy invasion as an intended feature rather than an exceptional, consent-bound action.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal