Pediatrics-PMPH-10edition

• Outcome: Review
• Latest audit: Jun 5, 2026 at 12:19 PM UTC
• Version:
  1.0.1

Overview

This is a content-only pediatrics reference skill, but it contains direct high-risk treatment, procedure, and dosing instructions that are not consistently gated to qualified clinical supervision.

Install only for qualified pediatric clinicians or supervised medical education. Do not use it for self-diagnosis, caregiver-directed treatment, emergency decision-making, or medication dosing without licensed clinician review, current local guidelines, and institution-specific protocols.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (52)

Intent-Code Divergence

High

• Confidence: 98% confidence
• Finding: The skill gives conflicting sleep-positioning guidance: it says all children must sleep in the left lateral position, but later states infants must never sleep prone due to SIDS risk. In a pediatric clinical guidance skill, contradictory positioning advice can directly mislead caregivers or clinicians, increasing risk of unsafe sleep practices, aspiration events, or delayed correction of harmful care.

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The skill gives concrete diagnostic provocations and treatment instructions for a rare pediatric metabolic disease, including OGTT dosing, glucagon stimulation testing, interpretation criteria, and acute glucose infusion rates, without clearly restricting use to qualified clinicians or emphasizing supervised, setting-specific decision making. In pediatric contexts, these actions can precipitate hypoglycemia or other harm if applied by non-experts or outside appropriate monitoring, so the operational framing materially increases safety risk.

Description-Behavior Mismatch

Medium

• Confidence: 95% confidence
• Finding: The skill is framed as an assessment aid, but its decision output instructs immediate treatment escalation ('立即实施灌肠复位' / '立即转手术治疗'). In a pediatric emergency context, converting an informational skill into an autonomous treatment directive is dangerous because it can bypass clinician judgment, omit confirmation by imaging/specialist evaluation, and contribute to delayed or inappropriate intervention if the input assessment is incomplete or wrong.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: This skill contains actionable pediatric diagnosis and treatment instructions for a high-risk condition, including steroid dosing, transfusion thresholds, immunosuppressants, plasmapheresis, and splenectomy criteria, but does not clearly state that it is for qualified clinician use only and must not be used as stand-alone medical advice. In the context of pediatric AIHA, omission of strong clinical-oversight warnings increases the chance of unsafe autonomous or layperson use, which could lead to misdiagnosis, delayed escalation, inappropriate transfusion, or harmful immunosuppressive treatment.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: The skill provides concrete anti-tuberculosis drug combinations, treatment durations, and pediatric meningitis dosing without an explicit warning that this is specialist medical guidance requiring clinician oversight, individualized diagnosis, and monitoring for toxicity, contraindications, and resistance. Because the content is directly actionable and concerns high-risk infectious disease treatment, an unqualified user or an agent could apply it inappropriately, causing delayed correct care, drug toxicity, undertreatment, or promotion of resistance.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill gives concrete triage and treatment guidance for pediatric acute abdominal emergencies, including medication class selection, surgical escalation, and contraindications, but does not clearly warn that this is reference material only and not a substitute for urgent clinician/surgical judgment. In a real-world agent setting, this omission could cause overreliance on the text during time-sensitive emergencies, delaying appropriate evaluation or leading to unsafe self-management.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The skill gives concrete neonatal clinical management guidance, including caloric targets, transfusion/EPO decisions, fluid restriction, sodium restriction, and diuretic use, but does not clearly state that it is not a substitute for neonatology or physician judgment. In this context, omission of an explicit warning is dangerous because BPD newborns are medically fragile and small errors in fluid, nutrition, or medication management can rapidly cause pulmonary edema, electrolyte disturbance, growth failure, or other serious harm.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: This skill gives clinically actionable risk stratification and surgical timing guidance for congenital heart disease without any warning that it is not a substitute for pediatric cardiology evaluation or urgent medical assessment. Because the topic involves cyanosis, pulmonary hypertension progression, Eisenmenger physiology, and operative contraindications, a user could over-rely on the skill and delay emergency referral or definitive specialist workup, creating a meaningful patient-safety risk.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill provides medical diagnostic guidance involving risk-bearing and invasive procedures, including contrast enema, anorectal manometry, and rectal biopsy, but does not explicitly require qualified clinical supervision or warn against layperson/self-use. In a medical knowledge skill, omission of supervision and safety boundaries can enable unsafe real-world action, delayed emergency care, or procedure-related harm, especially for neonates and infants.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The file gives a concrete NH₄Cl loading test protocol, dosing, monitoring intervals, and stopping thresholds, but it does not explicitly state that this is a clinician-supervised diagnostic procedure that must not be attempted by patients or lay users. In a generally accessible skill, procedural medical details can enable unsafe self-administration or unsupervised use, especially because ammonium chloride loading can worsen acidosis, hypokalemia, and other complications.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill ends by recommending escalation to steroids, immunoglobulin, or antibiotic changes without clearly restricting that advice to qualified clinicians or warning against self-directed treatment. In a pediatric context, these interventions can be high-risk and condition-dependent, so presenting them as a direct recommendation after a checklist-style diagnosis could encourage unsafe use or overconfidence by non-specialists.

Missing User Warnings

High

• Confidence: 96% confidence
• Finding: The skill provides concrete glucocorticoid initiation and maintenance dosing, adjustment thresholds, and monitoring steps for a serious pediatric neuromuscular disease, but it does not clearly warn that this is specialist-directed clinical reference material and not appropriate for unsupervised use. Because prednisone has significant risks including adrenal suppression, metabolic effects, infection risk, and dosing-related harm, presenting prescriptive treatment steps without a strong safety boundary can enable inappropriate self-management or misuse by non-specialists.

Missing User Warnings

High

• Confidence: 95% confidence
• Finding: The skill gives specific pediatric antiparasitic dosing, repeat-treatment timing, and household-wide treatment instructions without an explicit warning that medication choice, dosing, contraindications, age limits, pregnancy status of household contacts, and drug interactions require clinician or pharmacist confirmation. In this context, users may self-medicate children or family members based solely on the skill, creating a meaningful risk of dosing errors, inappropriate treatment of asymptomatic contacts, missed alternative diagnoses, or unsafe use in contraindicated individuals.

Missing User Warnings

High

• Confidence: 96% confidence
• Finding: The skill explicitly instructs users to perform a structured clinical diagnosis of epilepsy in suspected patients, including differential diagnosis, syndrome classification, etiologic workup, and functional assessment, but it does not state that the content is only educational support and not a substitute for licensed clinician judgment or urgent evaluation. In a medical context, especially for seizure-like events where delay or mis-triage can lead to serious harm, omission of safety boundaries can encourage overreliance on the skill and unsafe self-diagnosis or non-specialist diagnosis.

Missing User Warnings

High

• Confidence: 96% confidence
• Finding: The skill gives pediatric diagnostic conclusions and concrete treatment details, including drug dosing, without any warning that outputs must not replace clinician judgment or urgent in-person evaluation. In a pediatric infectious disease context, this is especially risky because misclassification, delayed escalation, or blindly followed dosing can cause serious harm in rapidly deteriorating children.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The skill explicitly offers IEI classification plus genetic testing and treatment-strategy suggestions, but it does not warn that the content is informational only or require clinician oversight. In a pediatric immunology context, users may over-trust the output for triage or management of suspected immune deficiency, creating risk of delayed diagnosis, inappropriate reassurance, or unsafe treatment decisions.

Missing User Warnings

High

• Confidence: 95% confidence
• Finding: The skill gives pediatric fluid-management guidance such as isotonic, hypertonic, and hypotonic replacement principles without an explicit warning that this is for licensed clinicians and not for layperson self-treatment. In a medical context, incomplete guardrails around rehydration advice can lead to dangerous misuse, including incorrect sodium correction, cerebral edema, seizures, or undertreatment of shock, especially in children.

Missing User Warnings

High

• Confidence: 95% confidence
• Finding: This skill provides treatment-selection guidance for diagnosed genetic diseases, including gene therapy, enzyme replacement, surgery, transplantation, drug therapy, and palliative care, but does not explicitly warn that it is not a substitute for licensed clinician judgment. Because it is framed as a workflow for choosing therapies, a user could over-rely on it to make or prioritize high-stakes pediatric treatment decisions, creating material patient-safety risk from inappropriate, delayed, or oversimplified care.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill provides detailed pediatric endocrine diagnostic instructions involving invasive and clinically risky procedures, including insulin-induced hypoglycemia testing, IV infusions, drug administration, MRI, and a treatment recommendation, but it does not clearly state that these steps must only be performed by qualified clinicians in an appropriately monitored medical setting. Because the target population is children and the workflow includes medications and dynamic testing with potentially serious adverse effects, omission of supervision and safety warnings increases the risk of unsafe self-application or misuse by unqualified users.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: This skill provides direct pediatric factor-replacement dosing formulas, treatment intervals, and duration for bleeding and perioperative management without any warning that outputs require clinician verification and individualized assessment. In a high-risk medical context involving children, omission of such safeguards can lead to underdosing or overdosing, inappropriate product selection, and delayed escalation in life-threatening bleeds such as CNS or retroperitoneal hemorrhage.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill gives directive medical guidance to 'immediately start HLH-1994 treatment' and to consider empiric treatment before diagnostic criteria are fully met, but it does not include a strong warning that the output must not be used as a standalone basis for treatment decisions. In a pediatric critical-care context, this is dangerous because HLH therapy is high-risk and misapplication to sepsis, malignancy, or other mimics could cause severe harm from immunosuppression, chemotherapy exposure, or delayed correct treatment.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: This index exposes a large catalog of pediatric diagnostic and treatment skills, including emergency care, drug dosing, resuscitation, and disease-specific interventions, but provides no prominent warning that the content is informational and not a substitute for licensed clinical judgment or urgent medical evaluation. In a medical context—especially pediatrics and neonatal/emergency care—users may over-rely on the tool for real-world treatment decisions, increasing risk of delayed escalation, dosing errors, or unsafe self-management.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The trigger phrase at line 35 instructs the system to immediately start a diagnostic workflow whenever a patient presents LCH-related multisystem symptoms. In a conversational agent, this is overly broad and can cause the skill to activate from symptom discussion alone, even when no biopsy exists or the user is asking hypothetically, which is especially risky in a medical context because it may drive premature diagnostic conclusions.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill gives immediate diagnostic and treatment directives for severe measles complications, including ICU transfer and condition-specific therapy, without an explicit limitation that it is for licensed clinicians using patient-specific judgment and local protocols. In a high-acuity pediatric context, such unqualified procedural guidance can be over-trusted, misapplied, or used outside appropriate supervision, increasing the risk of harmful delays, incorrect treatment, or unsafe triage.

Missing User Warnings

High

• Confidence: 95% confidence
• Finding: The skill provides step-by-step emergency treatment guidance for a life-threatening condition, including medication administration, IVIG/plasma exchange, airway intervention, and mechanical ventilation, but does not include any limitation notice restricting use to licensed clinicians or requiring immediate emergency escalation. In an agent setting, this can be acted on out of context, increasing the risk of unsafe medical decision-making, delayed transfer to emergency care, or harmful misuse by unqualified users.

VirusTotal

61/61 vendors flagged this skill as clean.