Obstetrics And Gynecology PMPH 10edition

• Outcome: Review
• Latest audit: Jun 4, 2026, 1:48 PM
• Version:
  1.0.0

Overview

This appears to be a medical reference skill, but it includes high-risk clinical treatment, procedure, and dosing guidance without enough safety boundaries for general installation.

Install only for qualified obstetrics/gynecology, oncology, reproductive medicine, or supervised medical education use. Do not use it as standalone patient advice, prescribing guidance, emergency care instructions, or procedural training. If installed, restrict access, require clinician review, verify all regimens against current local guidelines and institutional protocols, and add explicit warnings for emergency symptoms, invasive procedures, chemotherapy, pregnancy termination, and sensitive sexual-health data.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (44)

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The registry enumerates many high-risk obstetric, gynecologic, surgical, and emergency management skills, including invasive procedures and urgent clinical decision points, without any visible safety disclaimer or scope limitation. In a medical agent context, this can encourage overreliance on the skill content by non-specialists or use outside appropriate supervision, increasing the risk of harmful real-world medical actions.

Missing User Warnings

High

• Confidence: 95% confidence
• Finding: The skill provides specific hormone names, doses, routes, and durations for evaluating amenorrhea, but it does not clearly require licensed clinician oversight or state that these are prescription-only interventions requiring individualized risk assessment. In a medical skill, this omission is dangerous because users may self-administer estrogen/progestogen despite contraindications such as pregnancy, thromboembolic risk, liver disease, or structural pathology, leading to significant patient harm or delayed diagnosis.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: This is a true vulnerability because the skill gives directive, high-risk medical decision support for a neonatal endocrine/emergency condition, including immediate glucocorticoid treatment and possible genital surgery, without prominent warnings that specialist confirmation and urgent pediatric endocrinology evaluation are required. In this context, omission of safety guardrails is dangerous: misdiagnosis or premature treatment could cause inappropriate steroid exposure, missed salt-wasting crisis management, or irreversible surgical decisions.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: The skill gives stepwise guidance for diagnosing and treating a life-threatening obstetric condition, including drug dosing, intracervical/intralesional injection, curettage, vascular control, and hysterectomy escalation, but does not explicitly restrict use to qualified clinicians in an appropriately equipped medical setting. In a general agent environment, this omission is dangerous because it can normalize immediate invasive action by unqualified users or outside emergency-capable facilities, increasing the risk of hemorrhage, loss of fertility, or death.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: This skill gives concrete neonatal prophylaxis and treatment instructions for a high-risk medical context without any requirement to verify against current guidelines, clinician judgment, or specialist oversight. In addition, some listed recommendations appear clinically outdated or inappropriate, which increases the chance of unsafe care if an agent or user follows the content verbatim.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: The skill gives concrete pregnancy-related dosing instructions for vitamin B6, elemental iron, calcium, and laxative use without clearly requiring clinician supervision, dosage individualization, contraindication screening, or warning against self-medication during pregnancy. In an obstetrics context, users may act on this guidance directly, which can delay evaluation of serious conditions or cause inappropriate supplementation or medication use in a vulnerable population.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: This skill gives high-stakes clinical decision support, including statements about whether pregnancy is permissible and when pregnancy should be terminated, but it does not clearly state that it is informational only and cannot replace specialist cardiology-obstetric judgment. In this context, hard-coded rule-based recommendations may be over-applied by a user without considering patient-specific factors, creating a material risk of unsafe medical decisions or inappropriate pregnancy termination guidance.

Missing User Warnings

High

• Confidence: 96% confidence
• Finding: The skill provides concrete postoperative risk stratification and adjuvant treatment recommendations, including radiation, chemotherapy, and possible immunotherapy, without a clear user-facing warning that it is decision support rather than a substitute for oncology/gynecology specialist judgment. In a medical context, omission of such guardrails can lead users to over-trust the output and act on incomplete or non-individualized advice, creating a real risk of harmful undertreatment or overtreatment.

Missing User Warnings

High

• Confidence: 96% confidence
• Finding: This skill gives concrete medication, surgical indication, fertility, and cancer-triage recommendations for endometriosis without an explicit warning that outputs are informational and must not replace licensed clinician judgment. In a medical agent context, users may over-rely on the guidance for self-management or inappropriate triage, which could delay diagnosis, cause medication misuse, or lead to harmful decisions about surgery or fertility treatment.

Vague Triggers

Medium

• Confidence: 86% confidence
• Finding: The trigger examples are phrased as broad natural-language requests and do not define clear activation boundaries, exclusions, or disambiguation criteria. In an agent setting, this can cause the skill to activate on loosely related queries about textbook authority or expert qualifications, leading to unintended routing, over-application of this reference material, or interference with other skills' behavior.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill gives age-specific preventive and medical intervention guidance, including screening, hormone therapy, calcium supplementation, and management of gynecologic conditions, but does not warn that outputs are educational and not a substitute for clinician judgment or individualized care. In a medical domain, users may over-rely on generic guidance and delay professional evaluation, especially for symptoms like postmenopausal bleeding or decisions around vaccines, contraception, and hormone treatment.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: This skill provides explicit sexual-behavior techniques and multiple medication classes for treatment of female sexual dysfunction without clearly stating that these interventions require clinician supervision and individualized assessment. In this context, users could self-apply behavioral or drug guidance despite contraindications, trauma history, relationship coercion, or comorbid medical/psychiatric conditions, creating meaningful risk of physical and psychological harm.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The skill instructs collection of highly sensitive sexual-health, relationship, and psychological information without any user-facing privacy notice, consent guidance, or data-minimization safeguards. In an agent setting, this can lead to oversharing, improper retention, or transmission of intimate health data, creating privacy, compliance, and trust risks even if the clinical content itself is legitimate.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill provides explicit intimate clinical-style assessment guidance about adult female sexual response without any visible warning about sensitivity, privacy handling, informed consent, age gating, or the need for qualified clinical supervision. In a conversational agent context, this can lead to inappropriate use on non-clinical users, unsafe self-assessment, or mishandling of highly sensitive sexual health information.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: This skill gives step-by-step instructions for an invasive fetal surgical procedure, including anesthesia, trocar/fetoscope entry, vessel identification, laser coagulation strategy, and postoperative fluid reduction, without any warning that it is for trained maternal-fetal surgery specialists only. In an agent setting, such operational detail could enable unsafe reliance by unqualified users or be surfaced out of proper clinical context, creating a serious risk of maternal and fetal harm.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: The file provides specific chemotherapy agents, dosing, route, and cycle timing for gestational trophoblastic neoplasia without a clear, prominent warning that these are clinician-directed regimens requiring oncology supervision. Because the content concerns high-toxicity antineoplastic drugs with narrow therapeutic windows, misuse, self-administration attempts, or application outside appropriate specialist context could cause severe harm, including myelosuppression, organ toxicity, or death.

Missing User Warnings

High

• Confidence: 93% confidence
• Finding: This skill provides concrete ART regimen selection instructions for HIV-positive pregnant patients, a high-risk population where treatment choice depends on resistance history, comorbidities, gestational age, drug interactions, renal/hepatic function, and current guideline updates. Without an explicit warning that the content is not a substitute for specialist clinical judgment, the skill could be relied on as standalone treatment advice, increasing the risk of maternal harm, fetal harm, or failed prevention of vertical transmission.

Missing User Warnings

High

• Confidence: 98% confidence
• Finding: This skill provides diagnostic criteria, inpatient-level management, IV fluid support, complication screening, and prescription antiemetic/steroid treatment without a clear user-facing warning that hyperemesis gravidarum requires prompt clinician supervision and possible emergency care. In a pregnancy context, self-directed use could delay urgent evaluation, miss alternative serious diagnoses, or lead to unsafe medication use, electrolyte complications, or Wernicke encephalopathy.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: This index advertises a large catalog of high-risk obstetric and gynecologic diagnostic and treatment skills, including emergency management, invasive procedures, surgery, abortion, chemoradiation, fetal intervention, and drug protocols, but provides no visible safety disclaimer, clinician-only restriction, or emergency-care escalation warning. In a medical skill context, the absence of these guardrails increases the chance that non-professionals or underqualified users could treat the content as actionable clinical instructions, leading to delayed emergency care or unsafe self-management.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill gives a specific 'fetal hypoxia risk' alert and advises immediate medical follow-up based on a simplified self-screening threshold, but the warning that this is only preliminary screening is not sufficiently prominent at the point of output. In a pregnancy context, users may overinterpret the alert as a diagnosis or rely on the rule without considering urgent symptoms, which can cause inappropriate panic or delayed emergency care.

Missing User Warnings

High

• Confidence: 97% confidence
• Finding: The skill recommends clinician-directed treatments such as oxytocin, antibiotics, and ultrasound escalation without an explicit warning that the guidance is not a substitute for professional medical evaluation and prescribing. In a postpartum context, users may act on this advice during hemorrhage or infection, creating a significant risk of delayed care, inappropriate self-treatment, or harmful medication use.

Missing User Warnings

High

• Confidence: 98% confidence
• Finding: This file provides specific chemotherapy regimens and doses for ovarian cancer without any visible warning that the content is a reference only and must be verified against the patient's diagnosis, body surface area, renal function, prior toxicities, protocol version, and current oncology/pharmacy guidance. In a clinical skill context, omission of that warning increases the risk that a user could treat the table as executable dosing instruction, leading to serious medication error, toxicity, or undertreatment.

Vague Triggers

Medium

• Confidence: 86% confidence
• Finding: The trigger keywords are broad clinical phrases such as '保留生育功能', '年轻患者', and '术后化疗方案', which can appear in general medical discussion and may cause the skill to activate outside the narrow intended context. In a medical skill, mis-triggering is more dangerous because it can surface highly specific oncology treatment guidance in the wrong patient scenario, increasing the risk of inappropriate or overconfident recommendations.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill provides symptom-based diagnostic triage and suggests likely causes including serious conditions such as cervical cancer, endometrial cancer, and tubal cancer, but it does not clearly warn that the output is not a diagnosis and that urgent clinician evaluation may be required. In a medical context, users may over-rely on the skill, delay care, or misinterpret probabilistic associations as definitive conclusions, which can cause significant harm.

Natural-Language Policy Violations

Medium

• Confidence: 95% confidence
• Finding: The skill is explicitly tied to the 2011 China industry standard, but the metadata and workflow do not clearly warn that this is jurisdiction-specific guidance requiring regional applicability and user confirmation. In a medical diagnostic skill, applying a country-specific standard without explicit scoping can lead to inappropriate diagnosis when used in other regions or against newer guidelines, creating patient safety and clinical governance risk.

VirusTotal

64/64 vendors flagged this skill as clean.