Fix Your Entire Life in 1 Day

• Outcome: Pass
• Latest audit: Jul 10, 2026 · 2:24 AM UTC
• Version: 1.0.0

Overview

This is a local self-help journaling skill with no evidence of exfiltration or system persistence, though its prompts and saved reflections are sensitive.

Install only if you are comfortable with a forceful self-reflection style and with personal journal-like answers being saved in your local workspace. Do not treat it as therapy or crisis support, and review or delete the saved files if you do not want those reflections retained.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• 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
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (17)

Context-Inappropriate Capability

Medium

• Confidence: 90% confidence
• Finding: The skill asks whether it should create cron reminders, which crosses from coaching content into system-level task scheduling without a clear need tied to the skill’s core purpose. If the agent is connected to an environment where it can actually create scheduled jobs, this could cause unintended persistence, misconfiguration, or abuse of host resources beyond the user’s expectations.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill explicitly stores highly sensitive reflective content such as personal responses, psychological insights, and a final synthesized document under the user's workspace, but the user-facing documentation does not disclose retention, local storage, or privacy risks. In a psychology/self-improvement context, this data can contain intimate mental-health, identity, and life-history details, so undisclosed persistence increases the risk of unintended exposure to other tools, users, backups, or future compromise of the workspace.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: This session explicitly steers the user into vivid, regret-heavy future projection, social loss, and end-of-life reflection while framing distress as desirable (for example, 'Painful? Good.'). In a self-help skill with no screening, warning, grounding, or crisis guidance, this can aggravate anxiety, depression, shame, rumination, or hopelessness in vulnerable users.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: This session uses intentionally intense coaching language aimed at amplifying psychological discomfort ('make denial impossible,' 'staying the same becomes unbearable') without any upfront warning, opt-out framing, or emotional safety guidance. In a self-help skill, that can distress vulnerable users, trigger shame spirals, or push users toward unsafe emotional states, especially because the session explicitly seeks to increase tension rather than merely prompt reflection.

Natural-Language Policy Violations

Medium

• Confidence: 96% confidence
• Finding: The session explicitly frames being stuck or quitting as a marker of 'low intelligence,' which is shaming and derogatory language in a self-improvement context. In a psychologically oriented coaching skill, this can worsen self-criticism, discourage vulnerable users, and push them toward compliance through humiliation rather than supportive reflection.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: This session conducts a quasi-psychological assessment of the user's developmental stage and repeatedly solicits intimate beliefs, regressions, family-approval dynamics, and internal conflicts without any safety framing, privacy warning, or statement that it is not a clinical tool. That creates a real risk of emotional distress, overinterpretation of a non-validated framework, and unnecessary disclosure of sensitive personal information to the system.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The session frames the exercise as 'psychological surgery' and pushes intensive self-excavation around fear, shame, identity, and distress without any safety disclaimer, suitability screening, or escalation guidance. In a vulnerable user, this can intensify rumination, trigger acute distress, or substitute an unbounded AI exercise for appropriate mental health support.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The session explicitly guides the user into vivid, emotionally charged negative future visualization using coercive framing such as 'жестоко честен', hospice-regret authority cues, and prompts about end-of-life regret and shame, but provides no warning, consent check, grounding guidance, or crisis fallback. In a self-help skill aimed at broad users, this can trigger distress, rumination, hopelessness, or worsening symptoms in vulnerable users, especially those with anxiety, depression, trauma history, or suicidal ideation.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The session explicitly instructs the AI to infer users' 'hidden goals' and unconscious motives without any safety warning, limitation, or encouragement to treat outputs as speculative. In a self-help context, this can lead users to overtrust psychologically loaded interpretations, causing distress, shame, or misguided decisions based on unsupported inference.

Natural-Language Policy Violations

Medium

• Confidence: 97% confidence
• Finding: The file is entirely in Russian and does not present any language selection, fallback, or clear locale restriction. This can cause users or downstream agents to misunderstand the exercise, skip important safety-relevant context, or operate on content they cannot accurately review, which is a real quality and safety issue for a psychologically oriented skill.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: This is a real safety issue. The session explicitly encourages a 'burned bridge' commitment action and psychologically escalates pressure by making inaction feel intolerable, but it provides no warning, boundary, or instruction to avoid irreversible, high-risk decisions. In a self-help context, that increases the chance a distressed or impulsive user takes harmful financial, relational, employment, or mental-health-related actions without reflection.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: This session performs quasi-psychological assessment by assigning users to ego-development stages and prescribing transition exercises, but it provides no warning that the material is not mental-health advice and may trigger distress, shame, or identity destabilization. The risk is heightened by authoritative framing ('I will assess your stage') and prompts about deeply held beliefs, regression, dissolution, and loss, which can be harmful for vulnerable users or those with active mental-health concerns.

Natural-Language Policy Violations

Medium

• Confidence: 95% confidence
• Finding: The file is entirely in Russian and does not offer a language selection or explicit opt-in, which can cause users to engage with guidance they do not fully understand. In a self-improvement skill, misunderstanding instructions can lead to confusion, poor consent, and reduced ability to safely evaluate psychologically framed advice.

Natural-Language Policy Violations

Medium

• Confidence: 95% confidence
• Finding: The skill is entirely in Russian and does not offer a language selection or confirm the user's preferred language before beginning a psychologically sensitive self-reflection exercise. In this context, language mismatch can cause misunderstanding of nuanced prompts, reduce informed consent, and increase the risk of harmful misinterpretation during belief-challenging content.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The session instructs the user to perform emotionally intense 'psychological surgery' and confront feared truths, but it does not warn about distress, dissociation, or the need to pause and seek support if overwhelmed. In a mental-health-adjacent self-intervention context, this omission can amplify harm for vulnerable users, especially those with trauma, anxiety, or depression.

Missing User Warnings

Low

• Confidence: 88% confidence
• Finding: The skill offers to create cron reminders on the user's behalf without clearly warning that this may schedule device or system-level reminders. This can surprise users, create unintended persistent actions, and reduce informed consent around automation.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The reset command irreversibly deletes all persisted session state, notes, insights, and reminder files without any confirmation, guardrail, or soft-delete path. In a user-facing wellness skill that stores reflective personal content, accidental or indirect invocation can cause meaningful data loss and privacy-impacting destruction of user-authored records.

VirusTotal

66/66 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.