ClawHub
Back to skill

Security audit

Militia Training (EN)

Security checks for vulnerabilities and agentic risk

Overview

This skill is disclosed as militia training, but it contains actionable real-world weapons and combat-planning prompts that should be reviewed before installation.

Install only for clearly lawful, authorized, educational or simulation use. Be aware that the skill may guide firearm training and tactical combat planning; organizations should restrict or remove patrol, raid, ambush, and live-operations content before exposing it broadly.

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
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • 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
Findings (7)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The file contains prompt templates that facilitate real-world military tactics, firearms training, patrol planning, and battlefield medical response without clear safety boundaries, legal/ethical constraints, or requirements for qualified supervision. In this skill context, the omission is particularly dangerous because the content is explicitly framed for militia and combat use, which can enable violent misuse or unsafe handling of weapons and hazardous scenarios.

Content

No source excerpt is available for this finding.

Memory Manipulation

High
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · 01-implement-method.md (reported line 76)May include surrounding context.

md
INTELLIGENCE AVAILABLE: [WHAT IS KNOWN ABOUT THE AREA]

Provide a complete patrol order:
1. MISSION: Clear statement of what to accomplish
2. EXECUTION: Scheme of maneuver (formation, movement technique), fire team assignments, actions at each phase, actions on enemy contact (primary and contingency)
3. SUPPORTING: Communication plan (signals, frequencies), indirect fire coordination, medical support plan (casualty evacuation route)
4. SERVICE SUPPORT: Ammunition load plan, water and food, special equipment

Intent-Code Divergence

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The document's safety section says real-world violence instruction should be declined, yet the rest of the file provides detailed operational guidance for shooting, patrol formations, assault/support roles, and raid/ambush tactics. This contradiction is dangerous because it can create a false appearance of compliance while still equipping users with actionable violent methods, undermining safeguards and making policy evasion easier.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The trigger conditions are broad and cover multiple military-use topics such as shooting training, combat tactics, reconnaissance, and anti-aircraft defense without clear scoping or safety boundaries. This can cause the skill to activate in contexts involving operational weapon use or tactical guidance, increasing the chance that the assistant provides harmful dual-use or violent capability assistance beyond benign educational intent.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This section contains actionable small-unit combat guidance, including raid, ambush, route selection, formations, actions on contact, suppressive fire, and disengagement. Even though the skill is framed as a training assistant, these are real-world offensive and tactical instructions that could directly facilitate violent wrongdoing rather than merely providing historical or abstract educational context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
97% confidence
Finding

The manifest sets language: en, and the changelog notes the skill was rewritten in English, but the file does not offer a user language choice or explain a justified regional/compliance requirement. This is a natural-language policy concern because it appears to force a specific language by default.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

SQP-3 covers natural-language policy violations such as forcing a specific language without user opt-in. This markdown file provides all instructions, commands, and labels exclusively in English and does not offer any alternative language option or note that English is required for a justified region-specific purpose.

Content

No source excerpt is available for this finding.