Back to skill

Security audit

unisound-pre-existing-review

Security checks for vulnerabilities and agentic risk

Overview

The skill matches its medical-claims review purpose, but it can send sensitive claim text and a bearer API key to a user-configurable network endpoint.

Install only if you trust the publisher and can control how it is run. Use the documented internal HTTPS API endpoint, avoid overriding --api-url in production, provide a narrowly scoped key, de-identify claim materials before submission, and store stdout or --output files as sensitive medical/claims data.

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)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/run.py:103
Finding

Arbitrary API Endpoint Can Receive Bearer Credentials and Sensitive Medical Records

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding

The skill advertises no explicit tool/permission scope, yet its described implementation supports file read, optional file write via --output, and outbound network access to an internal medical model API. Without declared restrictions, a caller or runtime may grant broader capabilities than intended, increasing the chance of sensitive claims data being read from arbitrary paths, written to disk, or transmitted externally without clear policy enforcement.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The natural-language description, headings, inputs, and output template are all specified exclusively in Chinese, including fixed output labels such as “关系类型” and “是否报销”. This indicates a language constraint without any stated user choice or opt-in, which matches the locale/language policy violation criterion.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The default system prompt is entirely in Chinese and strictly constrains the output wording to fixed Chinese phrases such as 关系类型 and 是否报销. This imposes a specific language/locale behavior without any user opt-in or explanation that the skill is intentionally limited to Chinese-language workflows.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The code sends raw user-provided medical record or claims text to an external HTTPS API via urlopen without any explicit consent flow, warning, redaction step, or policy gate. Because the skill processes highly sensitive health data, this creates a real confidentiality and compliance risk if operators or users do not understand that protected information leaves the local environment.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill writes the full question and model answer to stdout and optionally to an output file, which can include medical-record content and reimbursement conclusions. Persisting this data locally without warning, minimization, or access-control guidance increases the chance of accidental disclosure through shell history capture, logs, shared filesystems, or insecure file handling.

Content

No source excerpt is available for this finding.