Back to skill

Security audit

Competitive Analysis Report

Security checks for vulnerabilities and agentic risk

Overview

The bundled skill processes notes locally with limited access, but its reports require manual source checks.

Check every figure and source manually; the supplied YAML template does not populate the script's sections correctly. Use a fresh output folder because digest.md is overwritten. The optional Windows executable is unnecessary for the Python workflow and was not reviewed. The publisher's [repository](https://github.com/rapidworker1/Multiple-AI-models-Toolkit) describes a broader AI toolkit.

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
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (2)

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_write.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The report builder does not ensure the manifest-promised source log for every figure.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

L90–L100 copy matching notes verbatim into the report, including numerical claims, while allowing the Source log section to contain only a missing-notes placeholder. The guardrail printed at L102–L104 does not enforce sourcing, so the generated report can include unsupported revenue, funding, customer-count, or margin figures despite the manifest's explicit sourcing promise.

Content

No source excerpt is available for this finding.