T09 · Insecure Skill Coding Practices
Warning
- Location
prompts/prompt_2_patient_retention.md:14- Finding
Marketing Communications Generated Without Verified Patient Consent
- Content
View full analysis
- Remediation
View remediation
Security audit
Chiropractic Pt Marketing Kit
Security checks for vulnerabilities and agentic risk
Overview
This is a Markdown-only healthcare marketing prompt kit with a real consent-caution issue, but it does not access data, send messages, run code, or persist anything.
Install only if you are comfortable reviewing healthcare marketing copy before use. Before sending any SMS, email, review request, reactivation message, newsletter, or referral campaign, confirm documented consent for that channel, suppress opt-outs, and have the final materials reviewed for your state and practice rules.
Vulnerability Patterns
- Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
- 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
Findings (1)
Vulnerability Patterns
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
- MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 87% confidence
- Finding
This markdown file describes the skill purpose in very broad terms: generating a complete patient lifecycle communication system covering multiple functions. It does not define specific trigger phrases, activation boundaries, or negative examples, which could cause unintended invocation for general patient-communications requests.
- Content
Vague Triggers
Medium
- Category
- Not specified by scanner
- Confidence
- 83% confidence
- Finding
The core instruction asks to 'Generate a complete patient communication and retention system' without defining when this skill should be selected versus other healthcare writing or compliance skills. In a markdown skill description, this missing specificity increases overlap with ordinary requests for emails, scripts, reminders, or patient communications.
- Content
Missing User Warnings
Low
- Category
- Not specified by scanner
- Confidence
- 83% confidence
- Finding
The prompt instructs generation of review-request scripts over SMS and email using patient first names and outreach channels, but it does not explicitly warn the user to use only patients who have provided the required marketing consent and contact authorization under HIPAA/TCPA/CAN-SPAM. Although the prompt includes several compliance rules, the omission of a clear operational warning can lead users to generate and deploy outreach workflows that process patient contact data in a regulated context without adequate consent controls or privacy review.
- Content
Natural-Language Policy Violations
Low
- Category
- Not specified by scanner
- Confidence
- 76% confidence
- Finding
The usage notes reference Nevada-specific chiropractic and physical therapy guidance and Nevada-local SEO targets, which introduces a locale-specific constraint into the skill's instructions. The file does not clearly state that the skill is only for Nevada practices or require the user to opt into Nevada-specific compliance framing before these instructions are applied.
- Content
Intent-Code Divergence
Low
- Category
- Not specified by scanner
- Confidence
- 77% confidence
- Finding
The prompt repeatedly frames output around compliance and cautious health advertising, but the usage note's recommendation to target keywords for 'highest SEO value' and 'very low competition' introduces marketing optimization guidance that is not reflected in the main compliance-oriented instructions. This is a mild intent/documentation divergence because the note shifts emphasis rather than directly changing code behavior, but it does not create a strong security-relevant contradiction.
- Content