skills.rest
notque
notqueCOMMUNITY
Explore all skills by this author →
GitHub

security-threat-model

Generate structured security threat models for the vexjoy-agent toolkit.

415|44|Updated Mar 18, 2026
Search Tags:#prompt injection detection#security threat modeling#supply-chain audit#deny-list generation#learning db sanitization#attack surface scanning
One-click install
npx skills add https://github.com/notque/vexjoy-agent --skill security-threat-model-notque
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: security-threat-model
Source: https://github.com/notque/vexjoy-agent/tree/main/skills/research/security-threat-model
Command: npx skills add https://github.com/notque/vexjoy-agent --skill security-threat-model-notque

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill prevents security gatekeeping gaps by systematically mapping your toolkit’s attack surface, deriving deny-list controls, auditing for prompt/payload injection risks, and checking learning DB contamination.

Core Features & Use Cases

  • Attack surface inventory: Enumerates hooks, MCP servers, installed skills, and risky environment references to understand what can be reached or influenced.
  • Deny-list generation: Produces actionable, human-reviewable permission denials tailored to detected risky patterns and baseline credential/command protections.
  • Supply-chain injection audit: Scans toolkit code and configuration for CRITICAL injection indicators such as zero-width/bidi characters, hidden payload blocks, ANTHROPIC_BASE_URL overrides, and role/instruction hijacking phrases.
  • Learning DB sanitization: Flags potentially injected entries from high-risk origins and unsafe content patterns, using dry-run reporting by default.

Quick Start

Run the security-threat-model skill to generate a threat model report under the security/ directory and block completion on any CRITICAL findings.

Frequently Asked Questions about security-threat-model

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How does prompt injection detection work for local agent toolkits?▼

Prompt injection detection scans installed skills, hooks, and learning DB content for CRITICAL indicators like hidden payload blocks, zero-width bidi characters, and role hijacking phrases. It generates a structured threat model report under the security directory, blocking completion if critical risks are found.

What is the best way to generate a deny-list for risky agent hooks and environment references?▼

Generating a deny-list requires enumerating your attack surface, including risky environment references and installed skills, to produce actionable permission denials. This process derives human-reviewable policy controls tailored to detected risky patterns and baseline credential protections.

How do I audit my learning database for injected entries and unsafe content?▼

Learning DB sanitization flags potentially injected entries from high-risk origins and unsafe content patterns using dry-run reporting by default. This process identifies contaminated data within your local installation, producing a validated JSON artifact for review before any changes are applied.

Can I run attack surface scanning on a local vexjoy-agent installation?▼

Attack surface scanning can be run on a local installation to enumerate reachable hooks, MCP servers, and installed skills. It requires deterministic Python scripts to produce JSON artifacts and enforces validation gates, blocking the threat model synthesis if CRITICAL findings are detected.

Why does threat model generation block completion when CRITICAL injection risks are found?▼

Threat model generation blocks completion on CRITICAL findings to prevent security gatekeeping gaps and enforce artifact-validation gates. This strict blocking mechanism ensures that severe supply-chain injection risks and role hijacking vulnerabilities are resolved before the validated document is synthesized.

Related Skills

Explore More Software Engineering Skills→
UsernameTronUsernameTron

setup-explainer

Scan installed Claude Code extensions and describe the environment in plain English.

Community
Intermediate
roderikroderik
2

sarif-parsing

Parse SARIF 2.1.0 files to extract, deduplicate, and summarize findings.

Community
Basic
jackquelinunpredictable827jackquelinunpredictable827
1

bidi-guard

Detect hidden Unicode BiDi characters that enable Trojan Source attacks.

Community
Intermediate
harperaaharperaa
15

security-prompts-threat-modeling

Generate STRIDE threat-model prompts for architecture, code reviews, and compliance assessments.

Community
Intermediate