T08 · Insecure Dependencies
Warning
- Location
icp_modeler.py:410- Finding
Dynamic Execution of an Undeclared External Python Module
- Content
View full analysis
- Remediation
View remediation
Security audit
ICP Modeler
Security checks for vulnerabilities and agentic risk
Overview
The skill mostly matches its ICP and marketing purpose, but its premium content feature can load undeclared Python code from local or external paths.
The offline ICP/profile features are straightforward. Treat the premium --generate-content path cautiously: only use it in an environment where you trust the generate.py module being imported, limit exposed API keys, and review all mortgage, VA-loan, credit-repair, or real-estate ad copy for accuracy and legal/compliance requirements before publishing.
Vulnerability Patterns
- Insecure DependenciesIntroduces malicious components through unsafe dependency sources
- 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
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
- MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)
Description-Behavior Mismatch
Medium
- Category
- Not specified by scanner
- Confidence
- 94% confidence
- Finding
The skill metadata and primary library suggest ICP profiling, but this function adds LLM-powered bespoke marketing-content generation. That expands the capability from analysis into content production, increasing the chance the skill is used to generate persuasive financial advertising or regulated claims without appropriate controls, review gates, or disclosure. In a mortgage/real-estate context, generated copy can easily become misleading or noncompliant.
- Content
Context-Inappropriate Capability
Medium
- Category
- Not specified by scanner
- Confidence
- 91% confidence
- Finding
The CLI exposes a generic '--generate-content' interface that accepts arbitrary requests, which is broader than narrowly modeling ICPs. Even though the prompt frames outputs around an ICP, this still grants an unrestricted marketing-copy generation feature that could be repurposed for misleading, manipulative, or noncompliant financial promotion. The risk is amplified because the domain is mortgage/credit advertising, where accuracy and regulatory compliance matter.
- Content
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 92% confidence
- Finding
LLM-generated marketing content is printed directly to the user without any warning that it may be inaccurate, misleading, or require legal/compliance review before publication. In the mortgage, VA-loan, and credit-repair space, outputs may include unsupported financial claims, outdated program details, fair-lending issues, or deceptive advertising language, making silent presentation particularly risky.
- Content