ClawHub
Back to skill

Security audit

Medical Billing Coder

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed medical billing assistance skill with healthcare privacy cautions and no hidden execution or persistence behavior.

Install only if you are comfortable using an advisory medical billing helper. Do not enter names, MRNs, DOBs, addresses, or other patient identifiers, and have a qualified coder or clinician review any suggested codes before billing. If you configure CMS_API_KEY, treat CMS validation as network egress and confirm your organization permits sending billing code pairs to CMS APIs.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · SKILL.md (reported line 4)May include surrounding context.

md
---
name: medical-billing-coder
version: 1.0.3
description: "Use this skill when a clinician, biller, or practice manager needs to look up ICD-10 diagnosis codes, CPT procedure codes, or E&M visit level codes. Takes a clinical note, visit summary, or plain-language description and suggests the most appropriate billing codes with confidence scores and audit trail. Also validates code combinations, checks for common denials, and flags upcoding or undercoding risks. DO NOT use as the sole basis for claim submission — all codes must be reviewed and approved by a qualified medical coder or clinician."
tags: ["healthcare", "billing", "icd-10",

External Transmission

Medium
Category
Data Exfiltration
Confidence
88% confidence
Finding

The skill declares optional external transmission to the CMS API, which creates a data egress path from a healthcare-adjacent workflow. Although the manifest says only billing code pairs are sent and no clinical text leaves the environment, this still introduces privacy, compliance, and integrity risk if the implementation ever sends more than intended or if users misunderstand what is transmitted.

Content

Scanner excerpt · SKILL.md (reported line 11)May include surrounding context.

md
homepage: https://github.com/optimusprime19/medical-billing-coder
repository: https://github.com/optimusprime19/medical-billing-coder
optionalEnv:
  - CMS_API_KEY  # Optional: calls https://api.cms.gov/medicare-coverage-database/v2 for real-time NCCI edit validation. No clinical text is transmitted — only code pairs are sent for validation.
---

# Medical Billing Code Suggester

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger phrases are broad enough that ordinary user requests like 'What CPT code for...' or 'Generate a superbill for this visit' could invoke the skill without clear confirmation that the user intends billing assistance or has provided sufficiently constrained, de-identified input. In a healthcare context, unintended invocation increases the chance of processing sensitive clinical content or producing coding guidance outside an appropriate review workflow.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.