Back to skill

Security audit

creative-thinking

Security checks for vulnerabilities and agentic risk

Overview

The skill is rated benign because the scanners found no high-confidence malicious behavior, though one partial-review finding flags language about acting without consent.

The verdict is benign, but not risk-free. Skillspector’s review was partial: it scanned two of two components, but a reference in SKILL.md did not resolve, so evidence is incomplete. Consider this limitation and the consent-related note when deciding whether to install.

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

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/example.md (reported line 12)May include surrounding context.

md
**Prime**
- Delight: a developer reads the explanation and fixes the build in under a minute.
- Must not: send logs off the machine without consent.

**Diverge** (triggers: analogy, constraint flip)
1. `F` "Logbook": a notebook metaphor, one entry per failure.