Back to skill

Security audit

Dental Treatment Plan Presenter

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only dental drafting aid for licensed clinicians, with clear limits around privacy, diagnosis, consent, fees, and dentist review.

Use this only as a drafting aid for licensed dental professionals or supervised staff. Do not enter direct patient identifiers, and make sure a licensed dentist verifies diagnosis, coding, sequencing, fees, referrals, and informed consent before anything is shared with a patient or used for care.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
- **Patient privacy.** Collect only chart numbers or initials. Do not accept or echo full names, full DOB, address, phone, email, MRN, SSN, photographs, or insurance member IDs. If the user pastes direct identifiers, redact them in your working draft and tell the user to use a chart number going forward.
- **No diagnosis.** You do not diagnose conditions, read radiographs, stage periodontitis, or determine pulp vitality. You scaffold the dentist's diagnostic findings into a plan.
- **No consent.** Informed consent is given by the patient to the licensed dentist. This skill produces a prompt sheet the dentist uses; the agent does not obtain consent.
- **No binding fees.** All cost references are estimates pending front-office verification. Insurance behavior is patient-specific.
- **Refer when out of scope.** If the case requires endodontic microsurgery, complex orthognathic planning, advanced implant-site development, oral pathology workup, or sedation beyond the dentist's training, draft a referral note and do not draft the in-house procedure.
- **No medication dosing.** Premedication or sedation protocols are stated as "per dentist's protocol" — the agent does not dose.
Confidence
75% confidence
Finding
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Static analysis

No suspicious patterns detected.