MindForge

• Outcome: Review
• Latest audit: Sep 16, 2026 · 5:38 AM UTC
• Version: 1.0.0

Overview

This is a self-help skill with no code or system access, but it gives broad addiction, withdrawal, ADHD, and physical-health advice without adequate safety limits.

Review carefully before installing or using this skill for addiction, substance use, prescribed medication changes, ADHD, severe burnout, or significant distress. It should be treated as informal self-help content only, not medical or mental-health guidance; users should seek qualified professional support for withdrawal, diagnosis, medication, crisis symptoms, or safety concerns.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

Findings (3)

other

Error

• Location: SKILL.md:64

Finding

Unqualified Universal Withdrawal Guidance May Delay Medical Care

Content

View full analysis

Vulnerability Details

File Location: SKILL.md, lines 64–86
Vulnerability Type: other: Unsafe Medical Guidance
Risk Level: High

Vulnerable Content

markdown

**Pain during withdrawal = the gardener at work. Don't interrupt it.**

---

## The 30-Day Dopamine Reset Timeline

### Phase 1: Days 1-10 — The Trough

- Dopamine baseline crashes
- Anxiety, insomnia, intense cravings peak
- This is NOT your normal state — it's withdrawal

### Phase 2: Days 11-20 — The Flatline

- "Pain gremlins" retreating
- Emotional numbness, zero libido
- This is HEALING — don't "test" yourself

### Phase 3: Days 21-30+ — The Reset

- D2 receptors upregulate
- Sensory experiences return (music, food, connection feel good again)
- PFC regains control over limbic system

Technical Analysis

The Skill is declared for use with “addiction,” but this guidance does not distinguish low-risk behavioral habits from physiological dependence on alcohol, benzodiazepines, opioids, or other substances. It describes a fixed 30-day biological progression and instructs users not to interrupt withdrawal despite anxiety, insomnia, and intense cravings.

Withdrawal severity, duration, and medical risk vary substantially by substance, exposure, comorbidities, and concurrent medication use. Some withdrawal syndromes can produce seizures, delirium, severe dehydration, cardiovascular instability, or other emergencies. Framing symptoms categorically as healing can discourage appropriate clinical assessment.

This issue does not grant an attacker system privileges and does not compromise the host environment. Its scope is user safety: it can influence vulnerable users to continue an unsafe self-directed withdrawal process.

Attack Path

1. A user asks the Skill for help with an addiction or stopping a substance.
2. The Skill applies its generalized 30-day reset model without determining the substance, dose, duration of u ...[truncated 759 chars]

Remediation

View remediation

Remediation Suggestions

1. Limit the self-directed reset protocol explicitly to low-risk behavioral habits unless a qualified clinician supervises the process.
2. State prominently that users should not abruptly stop alcohol, benzodiazepines, opioids, or prescribed medication based on the Skill.
3. Require professional assessment before offering withdrawal guidance involving substances or medications.
4. Remove categorical phrases such as “This is HEALING” and “Don't interrupt it.”
5. Replace the fixed timeline with qualified language explaining that symptoms and recovery periods vary.
6. Add urgent escalation guidance for seizures, hallucinations, confusion, chest pain, breathing difficulty, severe vomiting, suicidal thoughts, or rapidly worsening symptoms.
7. Clarify that the Skill is educational and is not a substitute for diagnosis, treatment, or emergency care.

other

Warning

• Location: neuroscience-addiction.md:120

Finding

Unsupported ADHD and Neuroendocrine Claims May Encourage Treatment Avoidance

Content

View full analysis

Vulnerability Details

File Location: neuroscience-addiction.md, lines 120–136
Vulnerability Type: other: Unsafe Medical Guidance
Risk Level: Medium

Vulnerable Content

markdown

### Acquired ADHD

Many diagnosed with adult ADHD don't have a developmental disorder — they have a dopamine baseline destroyed by supranormal stimulation.

**Mechanism:** Like a rock concert — when you're used to 120dB, a library (30dB) feels like silence. Your brain has turned down the volume receptors to protect themselves.

**Recovery:** Abstinence restores the natural range. The "ADHD" disappears.

### Testosterone-Dopamine Synergy

| Hormone | Role | Analogy |
|---------|------|---------|
| Testosterone | Fuel (available) | Gas in tank |
| Dopamine | Engine/throttle | How well the engine runs |

**In addiction:** You have bullets (testosterone) but a rusted trigger (dopamine system). The gun won't fire.

**After reset:** Dopamine mechanism repaired → testosterone can finally be "fired" functionally. You feel "high testosterone" — actually just restored dopaminergic expression of testosterone.

Technical Analysis

The document broadly suggests that many adult ADHD diagnoses are instead the result of a dopamine baseline “destroyed” by stimulation and claims that abstinence makes the condition disappear. It also presents a simplified relationship among addiction, dopamine, and testosterone as a settled physiological mechanism.

ADHD is a clinical diagnosis requiring consideration of developmental history, impairment across settings, differential diagnoses, sleep, substance use, medication effects, and other medical or psychiatric conditions. The content supplies no evidence qualification, diagnostic limitations, or warning against changing treatment. The categorical recovery claim may be interpreted as a diagnosis or cure.

This is not a technical privilege-escalation vulnerability. The relevant exploitation m ...[truncated 1012 chars]

Remediation

View remediation

Remediation Suggestions

1. Remove the assertion that many diagnosed adults do not have ADHD and the promise that abstinence makes ADHD disappear.
2. Clearly distinguish temporary attention difficulties associated with sleep, stress, or substance use from a clinical ADHD diagnosis.
3. State that only a qualified professional can assess ADHD and relevant differential diagnoses.
4. Warn users not to stop or modify prescribed medication without consulting the prescribing clinician.
5. Rewrite dopamine and testosterone discussions as limited, non-diagnostic educational material with appropriate uncertainty.
6. Cite reliable medical sources for material physiological claims and identify where evidence is preliminary or mixed.
7. Encourage evaluation when symptoms are persistent, severe, impairing, or accompanied by mood or safety concerns.

other

Warning

• Location: neuroscience-addiction.md:96

Finding

Maximum-Cold Shower Protocol Lacks Contraindications and Safety Controls

Content

View full analysis

Vulnerability Details

File Location: neuroscience-addiction.md, lines 96–102
Vulnerability Type: other: Unsafe Medical Guidance
Risk Level: Medium

Vulnerable Content

markdown

### 2. Cold Exposure

**Principle:** Cold water immersion spikes dopamine and norepinephrine 250%+, and the elevation persists for hours with NO subsequent crash.

**Protocol:** Last 1-2 minutes of shower at maximum cold.

**Key difference from addiction:** This dopamine is earned through discomfort — it strengthens the system rather than depleting it.

Technical Analysis

The Skill recommends one to two minutes at “maximum cold” without screening for cardiovascular disease, arrhythmias, uncontrolled blood pressure, respiratory conditions, pregnancy, fainting history, impaired temperature sensation, or other risk factors. It also provides no gradual acclimatization, supervision guidance, or stop conditions.

Sudden cold exposure can cause involuntary gasping, hyperventilation, abrupt cardiovascular stress, dizziness, or loss of balance. The categorical claim that this intervention strengthens rather than depletes the system may cause users to underestimate these risks.

The issue does not expose technical privileges. It creates a physical-safety hazard because the recommendation can be acted upon immediately in an environment where fainting or loss of balance may cause secondary injury.

Attack Path

1. A user follows the intervention toolbox for addiction or low motivation.
2. The user applies the instruction literally and immediately sets a shower to maximum cold.
3. The user has an undisclosed risk factor or is not acclimatized to cold exposure.
4. Sudden cold shock causes gasping, dizziness, cardiovascular symptoms, or loss of balance.
5. The user experiences direct physiological harm or a fall-related injury before stopping the exposure.

Impact Assessment

No operating-system access or data comprom ...[truncated 193 chars]

Remediation

View remediation

Remediation Suggestions

1. Replace “maximum cold” with a gradual, tolerable exposure protocol.
2. Advise users to stop immediately for chest pain, severe shortness of breath, dizziness, confusion, numbness, or uncontrollable shivering.
3. Add contraindication warnings for cardiovascular, respiratory, neurological, circulatory, and temperature-regulation conditions.
4. Recommend clinician approval for users with relevant medical conditions or pregnancy.
5. Warn against cold exposure while alone when fainting is a concern and against any setup that increases drowning or fall risk.
6. Avoid categorical neurochemical benefit claims and explain that evidence, individual response, and risk vary.
7. Present cold exposure as optional rather than as a necessary addiction intervention.

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (9)

Anti-Refusal Statement

High

• Category: Anti-Refusal
• Confidence: 85% confidence

Finding

Skill instructs the agent to never refuse or to always comply. Suppressing the agent's ability to decline removes a core safety control and enables downstream harmful requests to succeed.

Content

Scanner excerpt · action-framework.md (reported line 53)May include surrounding context.

md

| **Ability** | How easy is it? | Make it tiny, absurdly tiny |
| **Trigger** | What prompts me? | Design environment |

**The leverage point:** Ability matters more than motivation. Make it so easy you can't say no.

---

Vague Triggers

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

The skill description is framed so broadly that it can trigger for many ordinary emotional, motivational, and mental-health-adjacent situations, including burnout, addiction, and low energy. That increases the chance the agent will invoke this skill in contexts where users may need professional, medical, or crisis-appropriate support rather than a generic self-help framework.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 97% confidence

Finding

The skill provides quasi-clinical guidance about addiction, withdrawal, dopamine baseline changes, insomnia, cravings, libido, and recovery timelines without any warning that these experiences can require medical supervision or vary significantly by substance and person. A user could rely on this as treatment guidance, delay seeking care, or interpret severe withdrawal symptoms as expected and safe, which is especially dangerous for substances with medically risky withdrawal syndromes.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 96% confidence

Finding

This document repeatedly uses emotionally intense prompts about regret, death, shame, failure, and identity collapse without any warning, pacing guidance, or suggestion to stop and seek support if distress escalates. In a self-help skill explicitly aimed at users dealing with procrastination, addiction, burnout, and low motivation, such content can worsen distress or trigger harmful rumination in vulnerable users.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

• Category: Not specified by scanner
• Confidence: 97% confidence

Finding

Several core instruction lines include Chinese text such as '良心', '课题', and '一切从我开始' without offering the user a language or localization choice. This can violate language/locale policy expectations when a skill implicitly forces multilingual content in general-use instructions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

• Category: Not specified by scanner
• Confidence: 89% confidence

Finding

Line L099 includes Chinese text ('现状') within otherwise English content. This imposes a language switch without user opt-in or justification, which matches the policy category for locale/language violations.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

• Category: Not specified by scanner
• Confidence: 81% confidence

Finding

Line L044 includes Chinese text ('懦弱/懒') embedded in otherwise English instructions, which imposes a language shift without offering the user a choice or explaining the locale change. Under SQP-3, forced or unexplained language constraints can violate language/locale policy expectations when the skill otherwise appears intended for general English-language use.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

• Category: Not specified by scanner
• Confidence: 92% confidence

Finding

Line L112 includes the Chinese term "介入" embedded in otherwise English guidance, which introduces a language/locale inconsistency without offering the user a choice or documenting a justified locale-specific need. This matches the policy category for language or locale violations in natural-language content.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

• Category: Not specified by scanner
• Confidence: 83% confidence

Finding

This markdown file includes substantive instructional content in Chinese script (for example, quotations and terms) while the policy requires flagging language or locale constraints that are imposed without user opt-in. There is no indication that the user can choose a language or that the Chinese-only portions are required by a justified regional context.

Content

No source excerpt is available for this finding.