Back to skill

Security audit

Neurology-PMPH-9edition

Security checks for vulnerabilities and agentic risk

Overview

This is a medical neurology reference skill with no malware indicators, but it gives high-risk treatment and procedure guidance without consistent runtime safety boundaries.

Install only in a controlled medical education or clinician-review setting. Do not use this skill as standalone medical advice, for self-treatment, or for real-time emergency decisions unless a licensed clinician and current local protocols govern the output.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (118)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This skill provides procedural guidance for invasive, high-acuity stroke interventions including mechanical thrombectomy, intra-arterial thrombolysis, angioplasty, and stenting, but it does not clearly warn that the content is high-risk medical information requiring qualified specialist judgment. In this context, omission of a strong safety disclaimer is dangerous because users may over-rely on the output for real-time clinical decision-making in emergencies, potentially causing hemorrhage, delayed reperfusion, inappropriate patient selection, or other life-threatening harm.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill provides detailed diagnostic framing and treatment guidance for serious neurologic diseases such as ALS, MS, and autoimmune encephalitis, but it does not visibly warn that this content is for clinician reference only and not a substitute for urgent specialist evaluation. In a general-purpose agent environment, this can enable unsafe self-diagnosis or delayed emergency care for conditions involving seizures, respiratory compromise, or rapidly progressive neurologic decline.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This section recommends high-risk interventions including steroids, IVIG, plasma exchange, rituximab, cyclophosphamide, tocilizumab, intrathecal methotrexate, bortezomib, and tumor surgery without explicit safeguards about specialist oversight, contraindications, monitoring, or emergency setting requirements. If surfaced to an end user without strong controls, it could contribute to dangerous expectations, inappropriate requests for therapy, or delay in obtaining urgent inpatient neurologic and oncologic care.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This skill provides diagnostic criteria and frames the output as a basis for subsequent surgical intervention, but it does not warn that AI-generated interpretation is not a substitute for licensed radiologic or neurosurgical judgment. In a clinical context, this can encourage overreliance on the skill for diagnosis or operative decision-making, increasing risk of misdiagnosis, delayed specialist review, or inappropriate treatment.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill explicitly states that stenosis measurements will guide intervention decisions and balloon/stent sizing, but it does not include a clear warning that the output is only assistive and must not replace specialist clinical judgment. In this context, an incorrect measurement, wrong standard selection, or overreliance on the model could directly influence invasive cerebrovascular treatment planning and device selection, creating a significant patient-safety risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

This section gives acute anaphylaxis-style emergency treatment steps, including epinephrine administration and airway/circulatory management, without an explicit warning that it is clinician-facing guidance and not for lay self-treatment. In a general-purpose agent, users may follow these instructions outside an appropriate medical setting, causing delay in emergency care or unsafe drug administration.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill presents dialysis escalation and fluid/hemodynamic management as direct action steps without clearly stating these are specialist decisions requiring laboratory review, monitoring, and individualized assessment. A user could misinterpret the content as actionable self-management or non-specialist bedside direction, risking serious renal, electrolyte, or volume-related harm.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

This section provides post-stenting CHS blood-pressure targets and medication recommendations without clearly warning that new headache, seizures, vomiting, or neurologic deficits after cerebrovascular intervention require urgent emergency evaluation. Because these symptoms may indicate hemorrhage or hyperperfusion injury, self-adjusting blood pressure based on the text could delay imaging and definitive treatment.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

This skill provides concrete clinical targets, medication classes, dosing, contraindications, and decision logic for aspirin use in primary stroke prevention, but it does not present a clear user-facing warning that the content is medical guidance requiring clinician evaluation. In a medical skill, omission of such guardrails can lead users to self-medicate, misclassify risk, or apply recommendations outside appropriate clinical context, creating a real patient-safety hazard.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

This skill gives actionable medical treatment instructions, including sodium correction limits and interventions for suspected CPM, without an explicit warning that the content requires professional clinical judgment and close monitoring. Because hyponatremia correction errors can cause irreversible osmotic demyelination, omission of supervision and patient-safety warnings materially increases the risk of harmful real-world use by non-specialists or over-reliant agents.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill gives urgent clinical management recommendations, including lowering intracranial pressure and CSF diversion surgery, but does not clearly state that it is informational only and not a substitute for emergency evaluation by qualified pediatric/neurosurgical clinicians. In a high-acuity pediatric scenario, users may over-rely on the skill's output, delay escalation, or apply interventions outside appropriate supervision, creating material patient-safety risk.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

This skill provides detailed medical diagnostic criteria, differential diagnosis logic, MRI interpretation rules, and steroid/plasma-exchange treatment regimens without any visible warning that the content is informational and not a substitute for licensed clinical judgment or emergency care. In a medical skill that may be used at point of care, omission of such guardrails increases the risk that users over-rely on the skill for real patient management, potentially causing delayed escalation, misdiagnosis, or harmful treatment decisions.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

This file provides concrete diagnostic and treatment guidance, including antiepileptic drug selection, contraindications, and status epilepticus timing thresholds, but it does not prominently warn that the content is not a substitute for urgent clinical evaluation or emergency care. In a high-risk medical domain, omission of such guardrails can lead users to self-manage seizures or delay emergency treatment, especially for status epilepticus.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

This skill provides actionable seizure first-aid and anti-seizure medication adjustment workflows, including tapering, switching, and monitoring thresholds, but does not clearly state that it is not a substitute for professional medical care or that medication changes require clinician supervision. In a high-risk medical context, users may act on the instructions directly, leading to delayed emergency care, unsafe medication changes, breakthrough seizures, or status epilepticus.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill gives procedural instructions for EEG provocation testing, including hyperventilation, photic stimulation, and drug-induced sleep, but does not clearly restrict use to trained clinicians or present a prominent safety warning. These maneuvers can provoke seizures or other adverse events, so presenting them as a general-purpose workflow increases the risk of unsafe real-world execution by non-clinicians or in uncontrolled settings.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

This skill provides detailed diagnostic criteria, triage-relevant thresholds, and medication/treatment instructions for headaches, stroke classification, blood pressure management, drug withdrawal, and anticoagulation-related prevention guidance without any visible user-facing safety disclaimer or escalation guidance. In a medical skill covering potentially emergent neurologic conditions, users may treat the content as direct care instructions, creating a substantial risk of delayed emergency evaluation, inappropriate self-management, or harmful medication decisions.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

This skill provides diagnostic criteria, medication selection, dosing, contraindications, and treatment pathways for headache disorders without a clear warning that it is not a substitute for clinician assessment or emergency evaluation. In a medical skill, omission of guardrails is dangerous because users may self-diagnose secondary headache, misuse prescription drugs, miss contraindications, or delay urgent care for red-flag symptoms.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill provides concrete treatment recommendations, drug dosing, and contraindications without an explicit warning that management must be confirmed by qualified clinicians and individualized to patient-specific factors. In a high-stakes medical context, users may treat this as actionable care guidance, creating risk of medication misuse, delayed emergency evaluation, or harm from applying disease-specific advice before diagnosis is confirmed.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

This skill provides concrete intracerebral hemorrhage surgical triage thresholds, timing, and procedure recommendations without an explicit warning that it is not a substitute for licensed neurosurgical judgment or emergency care. In a high-acuity context, users may over-rely on the workflow as authoritative, which could contribute to delayed escalation, inappropriate surgery decisions, or harmful non-surgical management.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

This skill provides concrete intervention triage and procedure-selection guidance for ruptured and unruptured intracranial aneurysms, including contraindication screening and recommended endovascular techniques, but does not prominently warn that the output cannot replace specialist clinical judgment or emergency care. In a high-acuity neurology/neurosurgery context, a user could over-rely on the model's recommendation and delay urgent in-person evaluation or receive unsafe procedure suggestions based on incomplete data.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This skill provides detailed emergency neurological assessment and triage content for coma, brainstem localization, herniation risk, and urgent management outputs, but does not prominently warn that it cannot replace immediate in-person emergency evaluation. In this context, omission of an urgent-care disclaimer is dangerous because users may delay calling emergency services or rely on the model for life-threatening decision-making.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This section gives step-by-step lumbar puncture instructions, contraindication assessment, and complication handling, but does not prominently state that lumbar puncture is an invasive, high-risk procedure requiring licensed clinicians, sterile technique, imaging review when indicated, and institutional supervision. In the medical context, omission of these safeguards is particularly dangerous because users may treat the instructions as actionable procedural guidance, risking brain herniation, bleeding, infection, neurologic injury, or delayed emergency care.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

This skill provides detailed neurological localization and diagnostic classification guidance but does not clearly warn users that the output is informational support only and not a substitute for in-person clinical evaluation. In a medical context, omission of this boundary can encourage overreliance, delayed emergency care, or unsafe self-diagnosis, especially for symptoms like dysphagia, facial weakness, gait disturbance, or cognitive decline that may reflect stroke or other urgent disease.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

This skill gives detailed diagnosis support and stepwise treatment instructions for a high-risk condition, including immunosuppressants, anticoagulation, antiepileptics, and intracranial pressure management, but does not include a clear user-facing warning that it is not a substitute for clinician judgment or emergency care. In an agent setting, this can enable unsafe overreliance, delayed escalation, or inappropriate medication decisions for critically ill patients.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

This section presents invasive and high-risk interventions, including neurodestructive procedures, toxin injection, and surgery, in a directive treatment ladder without an explicit warning that the content is for trained clinicians and not for self-management or delayed emergency care. In a generally accessible agent skill, that omission can lead users to misinterpret specialist-only interventions as actionable advice, especially for conditions where facial pain, palsy, or cranial neuropathies may reflect urgent underlying disease.

Content

No source excerpt is available for this finding.