Security audit

Sinan's Thinking Adapter v2

Security checks across malware telemetry and agentic risk

Overview

This is a personalized teaching-style skill, but it directs persistent tracking of a named user's reasoning patterns after every session.

Review this skill before installing because it is built around a named individual and asks the agent to keep adding observations about how that person thinks. It has no executable code, but the Evolution Log should be limited to explicit, user-approved preferences with clear review and deletion controls.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Description-Behavior Mismatch

Medium

Confidence: 96% confidence
Finding: The skill goes beyond formatting explanations and instructs persistent tracking of a named user's reasoning traits after every session. That creates unnecessary behavioral profiling and retention of user-specific data, which can expose sensitive inference data and normalize memory updates without clear consent or purpose limitation.

Context-Inappropriate Capability

Medium

Confidence: 97% confidence
Finding: The Evolution Log contains cumulative session-by-session notes about the user's questions, growth, and reasoning style, which is broader than needed for a presentation-style skill. This is dangerous because it creates persistent user profiling and can capture behavioral patterns that the user may not expect to be stored long term.

Natural-Language Policy Violations

Medium

Confidence: 84% confidence
Finding: The skill mandates a rigid communication format and language policy for all interactions without asking the user each time whether this style is wanted. While not a classic security flaw, it can override user autonomy and lead to unwanted behavior persistence, especially when combined with the profiling instructions elsewhere in the file.

Ssd 3

Medium

Confidence: 97% confidence
Finding: A standing instruction to record and retain the user's reasoning patterns after every session is a persistent data-retention behavior. This is risky because it enables accumulation of behavioral metadata over time, which can be privacy-invasive and unnecessary for the stated goal of improving explanations.

Ssd 3

Medium

Confidence: 98% confidence
Finding: The Evolution Log explicitly stores dated observations about the user's behavior and questioning across sessions, which is a form of cumulative profiling. In context, this is more dangerous because the skill is framed as a teaching adapter, so the persistence is not obviously necessary and may happen without meaningful user awareness.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.