Skip to content

GodModeAI2025/owasp-mas-ai-skill

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

OWASP MAS AI Skill

The complete OWASP Mobile Application Security standard as a Claude AI Skill – built for auditors, developers, architects, and compliance teams. Includes machine-readable agent output for AI pipelines and self-sync with the OWASP source repositories.

License: CC BY-SA 4.0 OWASP MASVS OWASP MASWE OWASP MASTG


What is this?

A Claude Skill that consolidates three OWASP Mobile Application Security repositories into a single, AI-ready knowledge package:

OWASP Source Focus What the Skill produces
OWASP MASVS WHAT to verify – 24 controls across 8 categories Audit checklists, requirements matrices, compliance evidence
OWASP MASWE WHAT to look for – 117+ weaknesses Weakness lookups, CWE mappings, correlation analysis
OWASP MASTG HOW to test – techniques, tools, test procedures Test guidance, tool references, platform-specific test steps

The Skill combines these three sources into an end-to-end audit workflow, from requirements definition through weakness detection to finished assessment reports.

Installation

Build a .skill file

The repo contains the skill/ directory with all files. To create a .skill file (ZIP archive) that can be uploaded directly to Claude:

# Clone the repo
git clone https://github.com/GodModeAI2025/owasp-mas-ai-skill.git
cd owasp-mas-ai-skill

# Build the .skill file (ZIP of the skill/ directory)
cd skill
zip -r ../owasp-mas.skill .
cd ..

Drag the resulting owasp-mas.skill file into any Claude chat – done.

Manual installation as Custom Skill

Alternatively, copy the skill/ directory directly:

cp -r skill/ /path/to/skills/user/owasp-mas/

Contents

skill/
├── SKILL.md                              (Orchestrator – 7 roles)
└── references/
    ├── masvs-controls.md                 (All 24 MASVS v2.1 controls)
    ├── maswe-catalog.md                  (117 MASWE weaknesses + test approaches)
    ├── mastg-testing.md                  (Testing guide: Android + iOS + 20+ tools)
    ├── secure-coding-android.md          (Kotlin/Java secure coding patterns)
    ├── secure-coding-ios.md              (Swift secure coding patterns)
    ├── threat-model-template.md          (STRIDE template with MASVS mapping)
    ├── report-templates.md               (5 report formats + CWE mapping)
    ├── agent-output.md                   (Machine-readable schemas for AI pipelines)
    └── updater-sync.md                   (Self-sync with OWASP GitHub repos)

The 7 Roles

Role Triggers Output
Auditor audit, checklist, pentest, IPA/APK analysis MASVS checklist, findings, risk assessment
Developer secure coding, code review, implementation Platform-specific code, anti-patterns
Architect threat model, architecture review, requirements STRIDE analysis, requirements matrix, SADRs
Reporter report, executive summary, compliance 5 report formats (DOCX/MD/PPTX)
Weakness Analyzer MASWE ID, weakness, vulnerability Weakness details, CWE, remediation
Updater update skill, check OWASP updates Delta report, regenerated references, new .skill ZIP
Agent Output machine-readable, YAML, JSON, for agent, pipeline Structured artifacts for AI agent consumption

Agent Output for AI Pipelines

On request, the Skill produces machine-readable YAML/JSON artifacts that downstream AI agents can process directly:

Schema Downstream Agent
finding – findings with evidence and remediation Ticket agent, remediation agent, CI/CD gate
compliance – MASVS matrix with verdict GRC system, compliance agent
threat_model – STRIDE analysis with risk ratings Architecture agent, risk management
weakness – MASWE lookup with detection + fix Knowledge agent, vulnerability management
secure_coding – code pattern with platform code Code generation agent, PR review bot

Every output ends with an agent_summary block: metrics, blocked_controls, release gate decision, prioritized next_agent_actions. 10 downstream triggers for agent orchestration. 3 modes: Agent only | Human only | Dual (prose + YAML artifact).

Self-Sync with OWASP Sources

The Skill checks the three OWASP source repositories for new versions on demand and updates itself:

"Check for OWASP updates"
→ Quick check: version delta without rebuild

"Update the skill"
→ Full rebuild: regenerate references, validate, output new .skill ZIP

Supports partial rebuild and cascading updates (MASWE change → also updates CWE mapping and agent output schemas).

Coverage

MASVS v2.1 – all 8 categories, all 24 controls

Category Controls Focus
MASVS-STORAGE 2 Secure data storage (data at rest)
MASVS-CRYPTO 2 Cryptographic operations and key management
MASVS-AUTH 3 Authentication and authorization
MASVS-NETWORK 2 Network communication (data in transit)
MASVS-PLATFORM 3 Platform interaction and IPC
MASVS-CODE 4 Code quality and secure build practices
MASVS-RESILIENCE 4 Reverse engineering and tampering
MASVS-PRIVACY 4 Privacy and data protection

Additional Mappings

  • 117 MASWE weaknesses with severity and CWE
  • MASWE → OWASP Mobile Top 10 (2024)
  • MASVS → GDPR / PCI DSS / BSI / ISO 27001
  • 20+ tool references (Frida, Objection, Burp Suite, MobSF, Ghidra, ...)

Examples

"Analyze this IPA for MASVS compliance."
"What is MASWE-0013?"
"How do I implement certificate pinning in Swift?"
"Create a threat model for our banking app."
"Generate a technical assessment report from these findings."
"Output the findings as agent output."
"Check for OWASP updates."

Relationship to OWASP Sources

This project is a derivative of the following OWASP projects:

Project Repository License
OWASP MASVS github.com/OWASP/masvs CC BY-SA 4.0
OWASP MASWE github.com/OWASP/maswe CC BY-SA 4.0
OWASP MASTG github.com/OWASP/mastg CC BY-SA 4.0

The OWASP content was not copied verbatim but transformed for use as a Claude AI Skill: restructured into the Claude Skill format, consolidated from three separate repos, enriched with platform-specific code examples (Kotlin/Swift), extended with report templates and regulatory mappings, made agent-ready with YAML schemas and downstream triggers, and equipped with self-sync capabilities.

This project does not replace the originals. It makes them usable in an AI-driven workflow.

License

Licensed under CC BY-SA 4.0.

  • ✅ Share, adapt, commercial use
  • ⚠️ Attribution required (OWASP as source)
  • ⚠️ ShareAlike (derivatives must use same license)
This work is derived from the OWASP Mobile Application Security (MAS) project:
- OWASP MASVS (https://github.com/OWASP/masvs) – CC BY-SA 4.0
- OWASP MASWE (https://github.com/OWASP/maswe) – CC BY-SA 4.0
- OWASP MASTG (https://github.com/OWASP/mastg) – CC BY-SA 4.0

OWASP is a trademark of the OWASP Foundation.
This project is not affiliated with, endorsed by, or sponsored by the OWASP Foundation.

Author

Mark Zimmermann – AI strategist in the energy transition. 20+ years enterprise.

Contributing

Issues and pull requests are welcome. In particular:

  • Updates for new MASVS/MASWE/MASTG versions
  • Additional secure coding patterns (Flutter, React Native, .NET MAUI)
  • New agent output schemas and downstream triggers
  • Additional regulatory mappings (HIPAA, SOC 2, NIS2)
  • Translations

About

Den gesamten OWASP Mobile Application Security Standard als KI-Skill für Claude – nutzbar von Auditoren, Entwicklern, Architekten und Prüfern.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages