Security audit

Chinese Medicine 10edition

Security checks for vulnerabilities and agentic risk

Overview

This is a medical knowledge skill with no apparent technical malware, but it contains actionable diagnosis, treatment, dosing, toxic-herb, and invasive-procedure guidance that needs careful review before use.

Install only if you intend to use it as an educational or clinician-reviewed TCM reference. Do not rely on it for self-diagnosis, prescriptions, dosing, emergency symptoms, pregnancy or pediatric issues, cancer care, chest pain, toxic herbs, acupuncture, moxibustion, cupping, or other procedures without qualified medical supervision.

A.I.G

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

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (390)

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The skill explicitly instructs the agent to generate individualized diagnosis-linked treatment plans, including treatment principles, formulas, and drug compositions, but provides no safety framing, no limitation to educational use, and no warning that this is not a substitute for licensed medical care. In a medical context, especially one involving patient-specific recommendations, this can mislead users into relying on unverified clinical advice and may cause delayed care, inappropriate self-treatment, or harmful herb/drug interactions.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill instructs the agent to recommend specific TCM formulas such as 乌头汤 and external therapies like acupuncture, cupping, and medicated washing without any safety screening, contraindication checks, dosing limits, or advice to seek a licensed clinician. In a medical skill, this can lead users to rely on potentially unsafe treatment suggestions for undiagnosed conditions or conditions requiring urgent conventional care, increasing the risk of harm from toxic herbs, delayed diagnosis, or inappropriate external treatment.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill’s content is about national clinical medicine textbook revision policy, while the enclosing skill pack is described as a TCM clinical knowledge base of core clinical skills. This mismatch can cause an agent to retrieve and apply irrelevant political/editorial instructions in place of expected domain knowledge, creating prompt-scope confusion and increasing the chance of unsafe or manipulated behavior.
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 diagnostic criteria for abdominal/genital pain and testicular symptoms without any warning that it is educational only and not a substitute for licensed medical evaluation. In this context, users may self-diagnose serious causes of scrotal or lower abdominal pain and delay urgent care, which is especially dangerous because some differential diagnoses in that body region can require immediate 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 provides treatment-oriented syndrome differentiation for delayed menstruation without an explicit safety gate to first rule out pregnancy, ectopic pregnancy, or other urgent gynecologic/endocrine causes. In this context, users may incorrectly self-manage a potentially serious condition, delaying appropriate evaluation while following herbal or TCM treatment guidance.
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 condition differentiation and treatment-selection guidance for an eye disorder, including syndrome classification and therapeutic principles, but does not warn users that it is not a substitute for professional medical evaluation. Because it concerns diagnosis-like triage and treatment advice for ocular symptoms, a user could rely on it inappropriately and delay care for serious eye disease that may present similarly.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The file gives concrete external-use oncology remedies involving highly toxic substances such as arsenic compounds (信石/As₂O₃, 红砒) and mercurial material (轻粉), including a lesion-specific application for cervical cancer, but it does not prominently state that these are dangerous historical/specialist practices requiring licensed oncology/TCM supervision and not for self-treatment. In a knowledge skill covering clinical methods, this can be acted on as procedural guidance, increasing the risk of poisoning, tissue injury, delayed evidence-based cancer care, and serious patient harm.
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 step-by-step instructions for invasive acupuncture needle manipulation, including reinforcing/reducing techniques, without any safety warning, contraindications, sterility requirements, anatomical cautions, or restriction to licensed clinicians. In this context, omission of professional-supervision language materially increases the risk of injury, infection, bleeding, pneumothorax, and unsafe self-treatment or use by unqualified operators.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill gives concrete instructions for administering an emetic formula, including dose escalation until vomiting occurs, but does not clearly warn about the substantial medical risks of induced vomiting such as aspiration, dehydration, electrolyte imbalance, airway compromise, and delayed access to appropriate care. In this context, the skill is especially dangerous because it is framed as actionable clinical guidance rather than historical/reference information, increasing the chance a user could attempt unsafe self-treatment or treat another person without proper supervision.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The skill gives treatment-directing guidance for insomnia and explicitly states that all types should be paired with sedative/calming drugs or therapies, yet it provides no warning to seek licensed medical evaluation, no contraindications, and no escalation guidance for serious causes of insomnia. In a healthcare context, this can steer an agent to offer actionable medical advice that may delay diagnosis of psychiatric, neurologic, substance-related, or medication-induced causes and may encourage unsafe sedative use.
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 concrete treatment steps for unstable angina, including specific antiplatelet, statin, acupuncture, moxibustion, and acupoint-application guidance, but it does not prominently require immediate clinician supervision or urgent emergency escalation for chest pain. Because unstable angina is an acute coronary syndrome with potential progression to myocardial infarction, operational treatment instructions in a skill can encourage unsafe self-management or delayed emergency care.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill gives condition-specific diagnostic criteria, syndrome differentiation, and treatment recommendations, including named formulas, without any clear warning that the content is not a substitute for licensed medical evaluation. In a gynecology context, this can delay appropriate care, cause users to self-manage serious underlying conditions, and is especially risky because the skill only partially mentions referral/escalation scenarios.
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 concrete medical treatment instructions, including ingredient composition, preparation, application media, dosing frequency, and modern disease examples, but does not clearly instruct users to seek evaluation and supervision from a qualified clinician before use. Because it targets potentially serious infectious, inflammatory, and wound-related conditions such as cellulitis-like illness, mastitis, abscess, diabetic foot ulcer, pressure sore, and snake-bite swelling, omission of professional-guidance warnings can lead to delayed diagnosis, inappropriate self-treatment, worsening infection, or failure to obtain urgent care.
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 guidance for acupuncture and bloodletting-like point-pricking ('点刺出血') and frames it as something to execute, but it does not clearly instruct users to seek a licensed clinician or urgent medical care for red-flag symptoms such as high fever, altered consciousness, severe throat swelling, or neurologic symptoms. In a medical skill, this omission is dangerous because users may attempt invasive treatment themselves or delay appropriate emergency evaluation, creating risk of infection, bleeding, injury, and missed serious disease.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: This skill gives condition-specific treatment guidance for acute abnormal uterine bleeding, including immediate hemostatic action and formula selection, but does not clearly warn that heavy bleeding may require urgent in-person medical evaluation. In this context, users may delay emergency care or self-manage a potentially life-threatening condition, increasing the risk of hemorrhage, hemodynamic instability, or missed serious underlying disease.
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 infant diagnostic guidance and explicitly says it should be used to replace pulse diagnosis in children under 3, but it does not warn that the output should not be relied on as a sole basis for care or that urgent symptoms require immediate qualified medical evaluation. In the context of infants, omission of escalation and clinician-warning language is especially dangerous because caregivers or agents may over-trust the skill and delay appropriate treatment in serious illness.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: This skill gives actionable outbreak-response and medical/public-health guidance such as identifying severe infectious disease, isolating patients, and implementing preventive interventions, but it does not warn that the content is not a substitute for licensed medical or public-health advice. In a high-stakes infectious-disease context, users may over-rely on TCM-framed guidance and delay appropriate diagnosis, reporting, infection control, or urgent treatment, increasing risk to both the individual and the public.
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 condition-based herbal selection, combinations, and contraindication advice for symptomatic users without a clear warning that cough, wheeze, hemoptysis, infant illness, or persistent symptoms require licensed medical evaluation. In a medical-treatment context, users may self-medicate, misclassify symptoms (for example cold vs heat vs dryness), or use potentially toxic or contraindicated herbs such as banxia or bitter apricot seed, creating a meaningful risk of harm.
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 concrete medicinal selection, compatibility rules, dosing details for ginseng, and contraindications, but does not clearly warn users that it is not a substitute for professional diagnosis or treatment. Because the content concerns bleeding, prolapse, insomnia, and herb use with contraindications, omission of a prominent safety disclaimer can lead users to self-treat potentially serious conditions and misuse potent substances.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill gives step-by-step instructions for direct and scarring moxibustion, including deliberate creation of burns, sterile suppuration, blister management, and permanent scarring, but it does not present prominent user-facing safety warnings or require clinician supervision. In a general-purpose agent skill, this can enable untrained users to perform a hazardous procedure that may cause burns, infection, delayed care, or permanent injury.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill provides concrete medical triage, dosage ranges, preparation steps, and condition-specific modifications for high-fever states, pneumonia, and diabetic hyperglycemia without any warning to seek licensed medical care or emergency evaluation. In this context, users could self-treat potentially serious conditions and delay appropriate diagnosis or urgent treatment, creating significant risk of harm.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 97% confidence
Finding: This skill gives condition-specific treatment guidance, syndrome differentiation, and named herbal formulas for stroke sequelae without a clear user-facing warning to seek evaluation from a licensed clinician or avoid self-treatment. Because the condition involves post-stroke patients with potentially serious neurologic deficits, users may rely on the content instead of obtaining appropriate medical assessment, delaying care or using inappropriate treatments.
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 concrete medical triage and treatment guidance, including syndrome differentiation, formula selection, medication adjustment, and referral logic, without any warning that it is informational only and not a substitute for qualified clinical care. In a medical context, users may rely on the instructions for self-diagnosis or treatment, creating risk of delayed care, misdiagnosis, or unsafe herbal use.
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 step-by-step guidance for constructing multi-herb medical formulas, including role assignment and contraindication considerations, but it does not warn that such outputs require licensed clinical oversight. In a medical context, omission of a clear safety disclaimer increases the risk that users may treat the generated prescription design as actionable treatment advice, potentially causing harm through misdiagnosis, interactions, toxicity, or inappropriate use.
Content: No source excerpt is available for this finding.

Missing User Warnings

High

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill explicitly instructs generation of individualized TCM treatment plans and says it 'must' be used when formulating or adjusting treatment, but it does not include a clear user-facing medical safety disclaimer or escalation guidance for urgent, severe, or uncertain conditions. In a health context, this can encourage users to rely on model-generated treatment advice without clinician oversight, increasing the risk of delayed care, contraindicated remedies, or inappropriate self-treatment.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.