ClawHub
Back to skill

Security audit

First Principles Decomposer

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only problem-solving skill with no code execution, persistence, credential use, or hidden data access, though users should notice its broad activation phrases and privacy-sensitive examples.

Before installing, be aware that this skill may activate on common problem-solving phrases and steer answers into a structured first-principles format. Treat examples involving children's photos, face recognition, enrollment data, or scheduling as conceptual only; any real implementation would need explicit consent, privacy controls, retention limits, and legal review.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
80% confidence
Finding

Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Content

Scanner excerpt · references/integrated-frameworks.md (reported line 32)May include surrounding context.

md
### The Five Stages
1. **Empathize** - Deeply understand user needs, pain, context
2. **Define** - Frame the real problem (often different from stated problem)
3. **Ideate** - Generate many solutions without judgment
4. **Prototype** - Build quick, cheap versions to test
5. **Test** - Learn from real user feedback, iterate

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The trigger list is unusually broad and includes common conversational phrases such as 'break this down', 'from scratch', and 'what are we assuming', which can appear in ordinary user requests unrelated to explicit skill invocation. This creates a real risk of unintended activation, causing the agent to apply this decomposition workflow when not requested and potentially altering behavior, response structure, or downstream tool/skill chaining unexpectedly.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The example proposes AI face recognition and automatic parent-specific photo delivery in a childcare context without acknowledging consent, biometric data handling, access controls, retention limits, or legal/privacy obligations. In a reusable skill example, this can normalize deployment of a sensitive biometric system without safeguards, increasing the chance that downstream users implement privacy-invasive or noncompliant features.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The manifest describes a skill focused on breaking problems down to fundamental truths and rebuilding from first principles. This file documents a broader 'combined approach' that adds separate methodologies—Stanford Design Thinking and MIT Systems Engineering—and presents an integrated multi-phase method, which materially expands the skill's stated scope.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

The entire framework is written as fixed English instructional content and does not indicate whether users may request another language or locale. Under the policy rule, natural-language guidance that implicitly forces a specific language without opt-in can be treated as a language-policy issue.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.