Back to skill

Security audit

red-queen-effect

Security checks for vulnerabilities and agentic risk

Overview

This business-strategy coaching skill is proportionate to its purpose, with a small privacy caveat about saving lessons from real use.

Keep confidential company details out of reusable skill notes. If saving lessons from a session, approve the additions and use generalized, anonymized wording.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

The skill instructs recording actual usage failures in its reusable file without warning users about retaining their information.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The instruction to add observed entries after each real use and paste the actual failure pattern encourages retaining user-derived information in a reusable skill without consent or sanitization requirements. Because this skill analyzes competitive investments, margins, and strategic failures, those entries could expose confidential business information to later users or repository readers if saved or shared; however, the file does not explicitly require uploading data or demonstrate that persistence occurs.

Content

No source excerpt is available for this finding.