ClawHub
Back to skill

Security audit

Skill Maker

Security checks for vulnerabilities and agentic risk

Overview

This skill is a guide for creating other skills and does not include hidden execution, persistence, credential access, or destructive behavior.

Installers should understand that this skill helps an agent create or modify skill files when asked. Review generated skills carefully, especially their activation descriptions and any scripts or dependencies the agent creates, because this package provides authoring guidance rather than enforcing strict safety boundaries itself.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (4)

Self-Modification

High
Category
Rogue Agent
Content
├─────────────────────────────────────────────────────────────┤
│  1. INTERPRET  → What capability does this skill need?     │
│  2. DESIGN    → Structure, resources, trigger conditions   │
│  3. FORGE     → Write SKILL.md, create resources          │
│  4. TEST      → Verify triggers, check quality            │
│  5. POLISH    → Refine based on testing                   │
└─────────────────────────────────────────────────────────────┘
Confidence
85% confidence
Finding
Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The trigger guidance suggests very broad activation phrases like 'I need to [action]' and 'Help me with [domain]', which can match ordinary user language across many unrelated tasks. In an agent skill system, overly broad triggers can cause unintended skill activation, leading to prompt/context hijacking of downstream behavior or execution of inappropriate workflows.

Vague Triggers

Medium
Confidence
94% confidence
Finding
The activation template says 'This skill triggers when user wants to [capability]' without defining constraints, prerequisites, or non-trigger cases. That ambiguity encourages authors to write permissive activation rules, increasing the chance that a skill activates outside its intended context and influences agent behavior unexpectedly.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The weather example uses broad trigger wording such as 'User asks weather' and 'User wants forecast,' modeling imprecise trigger design for skill authors. Because this file is a skill-construction guide, insecure examples can propagate into many future skills and create systemic over-triggering across an ecosystem.

Static analysis

No suspicious patterns detected.