Security audit

Numerology Fortune

Security checks across malware telemetry and agentic risk

Overview

This numerology skill is not malicious, but it saves and silently reuses a user’s full legal name and birth date without clear opt-in or fresh confirmation.

Review before installing. Use it only if you are comfortable with your full legal name, birth date, and derived numerology profile being saved locally for future readings and potentially reused from a related fortune-hub memory file. For one-time use, instruct the agent not to write MEMORY.md and delete any existing saved profile data.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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 (5)

Context-Inappropriate Capability

Medium

Confidence: 96% confidence
Finding: The skill instructs the agent to persist full legal name and birth date, which are sensitive personal identifiers, in MEMORY.md for later reuse. Storing and reusing this data exceeds what is necessary for a single numerology response and creates avoidable privacy risk if the workspace, logs, or related skills can access the file.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: The instruction to read from fortune-hub/MEMORY.md expands data access beyond this skill's atomic numerology function and enables cross-skill reuse of personal profile data without fresh consent. This broadens the trust boundary and increases the chance of unintended data exposure or secondary use of sensitive information.

Missing User Warnings

Medium

Confidence: 98% confidence
Finding: The skill directs storage of sensitive personal data in MEMORY.md without a clear warning that the information will persist and may be reused later. Users may provide birth date and legal name for one-time calculation without realizing the data will be retained, creating a consent and privacy transparency failure.

Ssd 3

Medium

Confidence: 97% confidence
Finding: The instructions tell the agent to reuse previously stored identity data 'without asking again,' which bypasses fresh confirmation for sensitive personal information. That can lead to stale, incorrect, or unexpectedly reused data being incorporated into outputs, and it normalizes silent processing of sensitive user data.

Ssd 3

Medium

Confidence: 96% confidence
Finding: The memory template stores both sensitive source data and derived profile summaries for future use, increasing the amount of personal information retained over time. Even if each item seems individually low-risk, together they create a durable personal profile that could be exposed to other skills, operators, or unintended contexts.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.