PsyClaw PSS-10

• Outcome: Review
• Latest audit: May 29, 2026 at 9:29 PM UTC
• Version:
  0.1.0

Overview

This PSS-10 stress-screening skill is mostly coherent, but it sends complete stress-assessment answers and results to a remote platform without enough privacy, consent, or minimization guidance.

Install only if you are comfortable with PSS-10 answers and results being uploaded to the configured platform. Review the JSON before submission, use a scoped API key, avoid submitting another person’s responses without consent, and prefer sending only the minimum fields the platform actually needs.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill explicitly instructs the agent to compute a mental-health-related self-report locally and then return the complete JSON to the platform, but it provides no user-facing notice, consent step, or data-minimization guidance. Because the payload includes answers, scores, classification, and severity, it can expose sensitive health-adjacent personal data to a remote service without adequate privacy disclosure.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The curl example demonstrates authenticated transmission of assessment results to a remote API but omits any warning that the content may contain sensitive self-reported stress or wellness information. This normalizes sending private assessment data off-device and could lead integrators to deploy the skill without informing users or implementing appropriate safeguards.

VirusTotal

63/63 vendors flagged this skill as clean.