Back to skill

Security audit

Test Time Compute Guide

Security checks for vulnerabilities and agentic risk

Overview

This is a small educational skill about test-time compute techniques, with no hidden execution, persistence, or data access behavior found.

Installers should treat this as an educational reasoning guide. When applying its examples, use reasoning formats allowed by your organization or platform and avoid asking models to disclose hidden internal reasoning if that is not permitted.

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
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
76% confidence
Finding

The package description and keywords explicitly promote "chain-of-thought reasoning techniques," which can imply exposing or requiring a specific internal reasoning style. Natural-language policy guidance often requires that such behavior not be forced or implied without clear user choice or boundary-setting, and this manifest provides no such clarification.

Content

No source excerpt is available for this finding.