Skip to content
Back to search

ai-security-arsenal

SkillWorks with Claude CodeGitHub

A collection of skills, agents, commands, and workflows for security researchers. Compatible with Claude Code, Claude Desktop, OpenCode, and other AI coding tools.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/hardw00t--ai-security-arsenal/badge.svg)](/listings/hardw00t--ai-security-arsenal)

HTML

<a href="/listings/hardw00t--ai-security-arsenal"><img src="/api/listings/hardw00t--ai-security-arsenal/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Mobile Testing

# Android - Connect rooted device and start Frida
adb devices
adb shell "su -c '/data/local/tmp/frida-server -D &'"
# "pentest com.example.app"

# iOS - Connect jailbroken device
iproxy 2222 22
ssh -p 2222 root@localhost
# "pentest com.example.app"

Web Testing

# DAST with Playwright MCP
# "scan https://target.com for vulnerabilities"

# API Security
# "test API security for https://api.target.com"

Cloud/Infrastructure

# Cloud Security
aws configure  # or gcloud auth login
# "assess AWS security for account 123456789"

# IaC Security
# "scan Terraform in ./infrastructure for security issues"

# Container Security
# "scan container image nginx:latest"

Code Analysis

# SAST
# "scan this codebase for vulnerabilities"

# SCA
# "check dependencies for vulnerabilities"

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 438 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # AI Security Arsenal
  2. A collection of skills for automated security testing, penetration testing, and security assessment workflows with coding agents like Claude Code, Codex, OpenCode, and others.
  3. Each skill is structured as a thin **router** (`SKILL.md`, 150-250 lines) with lazy-loaded sibling content (`workflows/`, `methodology/`, `payloads/`, `references/`, `examples/`, `schemas/`). This lets agents load only what a sub-task needs instead of parsing a monolithic document every turn — leaving more context for the actual target code, traffic, and tool output.
  4. **Last validated:** 2026-04 against frontier coding agents (Claude Opus 4.x, GPT-5.x class).
  5. ## Skills Overview
  6. | Category | Skill | Description |
  7. |----------|-------|-------------|
  8. | **Mobile** | [android-pentest](#android-pentest) | Android app pentesting with Frida, objection, ADB |
  9. | **Mobile** | [ios-pentest](#ios-pentest) | iOS app security testing with Frida, Cycript, objection |
  10. | **Web** | [dast-automation](#dast-automation) | Dynamic testing with ZAP, Burp, Nuclei, Playwright |
  11. | **Web** | [api-security](#api-security) | REST/GraphQL API security, OWASP API Top 10 |
  12. | **Cloud** | [cloud-security](#cloud-security) | AWS/GCP/Azure security assessment |
  13. | **Cloud** | [iac-security](#iac-security) | Infrastructure as Code security scanning |
  14. | **Cloud** | [container-security](#container-security) | Docker/Kubernetes security assessment |
  15. | **Network** | [network-pentest](#network-pentest) | Internal network/AD penetration testing |
  16. | **Code** | [sast-orchestration](#sast-orchestration) | Static analysis with Semgrep, CodeQL, Bandit |
  17. | **Code** | [sca-security](#sca-security) | Software Composition Analysis, SBOM, dependencies |
  18. | **AI/ML** | [llm-security](#llm-security) | LLM security, prompt injection, OWASP LLM Top 10 |
  19. | **Process** | [threat-modeling](#threat-modeling) | STRIDE, PASTA, Attack Trees, threat analysis |
  20. ---
  21. ## Skill Architecture
  22. Every skill follows the same layout so agents can navigate them predictably:
  23. ```
  24. skills/<skill-name>/
  25. ├── SKILL.md # Thin router (150-250 lines): triggers, decision tree, indexes
  26. ├── workflows/ # Step-by-step runbooks loaded per task
  27. ├── methodology/ # Deep methodology docs per phase
  28. ├── payloads/ # Payload/injection/query files loaded on demand
  29. ├── references/ # Static reference material (OWASP, CWE, tool docs)
  30. ├── examples/ # Concise tool-call blueprints and worked examples
  31. ├── templates/ # Report/finding templates
  32. ├── schemas/
  33. │ └── finding.json # JSON Schema for structured findings
  34. └── scripts/ # Executable helpers (where present)
  35. ```
  36. Every `SKILL.md` includes:
  37. - **When NOT to use** — explicit pointers to sibling skills for overlap cases
  38. - **Decision tree** — how to pick the right workflow
  39. - **Parallelism hints** — which operations run concurrently vs sequentially
  40. - **Sub-agent delegation** — when to spawn parallel sub-agents (e.g., one per domain, one per tool, one per auth context)
  41. - **Reasoning budget** — where extended thinking pays off vs where to execute without it
  42. - **Multimodal hooks** — screenshots, diagrams, image-based evidence (where relevant)
  43. - **Output schema reference** — points at `schemas/finding.json`
  44. Findings validate against the per-skill JSON Schema, so output can be piped directly into SARIF converters, Jira, DefectDojo, or custom triage tooling.
  45. ### Headline workflows (new in 2026-04)
  46. | Skill | Workflow | What it does |
  47. |---|---|---|
  48. | sast-orchestration | `workflows/triage.md` | SARIF ingest + exploitability ranking with extended thinking |
  49. | sca-security | `workflows/reachability_analysis.md` | Combines SBOM + SAST call graphs to filter out unreachable CVEs |
  50. | sca-security | `workflows/lockfile_diff.md` | PR-time review of new dependencies, new vulns, license drift |
  51. | container-security | `workflows/sbom_diff.md` | Compare prior vs current SBOM, flag newly-introduced CVEs |
  52. | threat-modeling | `workflows/stride_from_openapi.md` | Auto-generate STRIDE-per-interaction from an OpenAPI spec |
  53. | iac-security | `workflows/policy_as_code_loop.md` | Iterative OPA/Rego rule authoring against vulnerable fixtures |
  54. | llm-security | `workflows/agentic_tool_misuse.md` et al. | Agentic, memory, MCP, skill-file, computer-use, multimodal injection |
  55. ---
  56. ## Mobile Security
  57. ### Android Pentest
  58. Comprehensive Android mobile application penetration testing with full ADB shell access to rooted devices.
  59. **Capabilities:**
  60. - Static & dynamic analysis of Android applications
  61. - Runtime manipulation with Frida (SSL pinning bypass, root detection bypass, crypto hooks)
  62. - OWASP MASTG compliance testing
  63. - Data extraction (databases, SharedPreferences, storage)
  64. - Component fuzzing (activities, services, content providers, broadcast receivers)
  65. **Trigger phrases:** `pentest this Android app`, `bypass SSL pinning`, `extract data from app`, `MASTG testing`
  66. **Tools:** ADB, Frida, objection, apktool, jadx, Burp Suite
  67. ---
  68. ### iOS Pentest
  69. iOS mobile application security testing with runtime manipulation and binary analysis.
  70. **Capabilities:**
  71. - Static analysis with class-dump, otool, Hopper
  72. - Runtime manipulation with Frida and Cycript
  73. - SSL pinning and jailbreak detection bypass
  74. - Keychain extraction and data protection analysis
  75. - OWASP MASTG compliance testing
  76. - IPA decryption and binary patching
  77. **Trigger phrases:** `pentest this iOS app`, `bypass jailbreak detection`, `extract keychain`, `iOS security assessment`
  78. **Tools:** Frida, objection, Cycript, class-dump, otool, Clutch/bfdecrypt
  79. ---
  80. ### Mobile MCP for Device Interaction
  81. Both mobile pentesting skills support **[Mobile MCP](https://github.com/mobile-next/mobile-mcp)** for advanced device, emulator, and simulator interactions:
  82. - Screen capture and UI inspection
  83. - Touch automation and gesture simulation
  84. - App lifecycle management
  85. - Complements Frida/Objection for scenarios requiring direct UI interaction
  86. ```json
  87. // Add to ~/.claude/mcp.json
  88. {
  89. "mcpServers": {
  90. "mobile-mcp": {
  91. "command": "npx",
  92. "args": ["-y", "@anthropic/mobile-mcp", "--android"] // or "--ios"
  93. }
  94. }
  95. }
  96. ```
  97. ---
  98. ## Web Security
  99. ### DAST Automation
  100. Automated Dynamic Application Security Testing using browser-based scanning and traditional tools.
  101. **Capabilities:**
  102. - Blackbox and greybox (authenticated) testing
  103. - Intelligent browser-based crawling with JavaScript execution
  104. - Multi-domain parallel scanning
  105. - XSS, CSRF, SQLi, SSRF, IDOR, auth bypass detection
  106. - Integration with Nuclei, ZAP, Burp Suite, SQLMap
  107. **Trigger phrases:** `scan domain for vulnerabilities`, `run DAST on`, `automated security scan`, `web penetration test`
  108. **Tools:** Playwright MCP, Nuclei, OWASP ZAP, Burp Suite, SQLMap
  109. ---
  110. ### API Security
  111. REST and GraphQL API security testing based on OWASP API Security Top 10.
  112. **Capabilities:**
  113. - Authentication/authorization testing (BOLA, BFLA, broken auth)
  114. - Injection testing (SQLi, NoSQLi, command injection)
  115. - Mass assignment and SSRF detection
  116. - Rate limiting and resource exhaustion testing
  117. - GraphQL introspection and query depth attacks
  118. - JWT and OAuth vulnerability testing
  119. **Trigger phrases:** `test API security`, `scan REST API`, `GraphQL security assessment`, `OWASP API Top 10`
  120. **Tools:** Burp Suite, Postman, graphql-cop, jwt_tool, Nuclei
  121. ---
  122. ## Cloud Security
  123. ### Cloud Security
  124. Multi-cloud security assessment for AWS, GCP, and Azure environments.
  125. **Capabilities:**
  126. - IAM policy analysis and privilege escalation paths
  127. - Storage security (S3, GCS, Azure Blob misconfigurations)
  128. - Network security (VPC, security groups, firewall rules)
  129. - Secrets management and key rotation assessment
  130. - Compliance checking (CIS benchmarks, SOC2, PCI-DSS)
  131. - Serverless security (Lambda, Cloud Functions)
  132. **Trigger phrases:** `assess AWS security`, `audit GCP permissions`, `Azure security review`, `cloud misconfiguration scan`
  133. **Tools:** ScoutSuite, Prowler, CloudSploit, Steampipe, cloud-nuke
  134. ---
  135. ### IaC Security
  136. Infrastructure as Code security scanning for Terraform, CloudFormation, Kubernetes, and Helm.
  137. **Capabilities:**
  138. - Terraform security with Checkov, tfsec, Terrascan
  139. - CloudFormation scanning with cfn-lint, cfn-nag
  140. - Kubernetes manifest security with kubesec, kube-linter
  141. - Helm chart security analysis
  142. - ARM template scanning for Azure
  143. - Policy-as-code with OPA/Conftest
  144. **Trigger phrases:** `scan Terraform for security issues`, `check CloudFormation template`, `Kubernetes manifest security`, `IaC security audit`
  145. **Tools:** Checkov, tfsec, Terrascan, KICS, kubesec, OPA/Conftest
  146. ---
  147. ### Container Security
  148. Docker and Kubernetes security assessment including image scanning and cluster hardening.
  149. **Capabilities:**
  150. - Container image vulnerability scanning with Trivy, Grype
  151. - SBOM generation with Syft
  152. - Kubernetes cluster security with Kubescape, kube-bench
  153. - Runtime threat detection with Falco
  154. - Container escape and privilege escalation testing
  155. - Network policy and RBAC analysis
  156. **Trigger phrases:** `scan container image`, `Kubernetes security audit`, `check Docker security`, `cluster hardening assessment`
  157. **Tools:** Trivy, Grype, Syft, Kubescape, kube-bench, kube-hunter, Falco
  158. ---
  159. ## Network Security
  160. ### Network Pentest
  161. Internal network and Active Directory penetration testing.
  162. **Capabilities:**
  163. - Network reconnaissance with Nmap, Masscan
  164. - Active Directory enumeration with BloodHound
  165. - Credential attacks (Kerberoasting, AS-REP roasting, password spraying)
  166. - Lateral movement (Pass-the-Hash, Pass-the-Ticket, DCOM, WMI)
  167. - Privilege escalation (DCSync, delegation abuse, GPO attacks)
  168. - Domain dominance (Golden/Silver tickets, skeleton key)
  169. **Trigger phrases:** `internal network pentest`, `Active Directory assessment`, `enumerate domain`, `lateral movement testing`
  170. **Tools:** Nmap, BloodHound, Impacket, CrackMapExec, Responder, Mimikatz, Rubeus
  171. ---
  172. ## Code Security
  173. ### SAST Orchestration
  174. Static Application Security Testing orchestration with custom rule development.
  175. **Capabilities:**
  176. - Multi-tool scanning with Semgrep, CodeQL, Bandit
  177. - Custom rule development (Semgrep YAML, CodeQL queries)
  178. - Finding triage and false positive reduction
  179. - CI/CD integration (GitHub Actions, GitLab CI)
  180. - Language-specific scanning (gosec, Brakeman, ESLint security)
  181. - SARIF output and result aggregation
  182. **Trigger phrases:** `scan code for vulnerabilities`, `write Semgrep rule`, `triage SAST findings`, `static analysis setup`
  183. **Tools:** Semgrep, CodeQL, Bandit, gosec, Brakeman, SpotBugs, ESLint
  184. ---
  185. ### SCA Security
  186. Software Composition Analysis for dependency vulnerabilities and supply chain security.
  187. **Capabilities:**
  188. - SBOM generation (CycloneDX, SPDX formats)
  189. - Multi-ecosystem scanning (npm, pip, Maven, Cargo, Go, Ruby, PHP)
  190. - Vulnerability database correlation (NVD, OSV, GitHub Advisory)
  191. - License compliance checking
  192. - Dependency graph analysis and transitive vulnerability detection
  193. - Supply chain attack detection
  194. **Trigger phrases:** `scan dependencies`, `generate SBOM`, `check for vulnerable packages`, `license compliance audit`
  195. **Tools:** Syft, Grype, Trivy, OWASP Dependency-Check, npm audit, pip-audit, Snyk
  196. ---
  197. ## AI/ML Security
  198. ### LLM Security
  199. LLM and AI application security testing including prompt injection and guardrail testing.
  200. **Capabilities:**
  201. - Prompt injection testing (direct and indirect)
  202. - Jailbreaking and guardrail bypass techniques
  203. - System prompt extraction attacks
  204. - RAG pipeline security assessment
  205. - AI agent excessive agency testing
  206. - OWASP LLM Top 10 coverage
  207. **Trigger phrases:** `test LLM for prompt injection`, `jailbreak AI system`, `test AI guardrails`, `RAG security assessment`
  208. **Tools:** Custom scripts, Garak, AI security testing frameworks
  209. ---
  210. ## Security Process
  211. ### Threat Modeling
  212. Systematic threat identification using industry frameworks and methodologies.
  213. **Capabilities:**
  214. - STRIDE threat analysis (per element and per interaction)
  215. - PASTA 7-stage methodology
  216. - Attack Tree construction with AND/OR gates
  217. - DREAD risk scoring
  218. - Data Flow Diagram (DFD) creation
  219. - Threat library and mitigation mapping
  220. **Trigger phrases:** `threat model this system`, `STRIDE analysis`, `create attack tree`, `security architecture review`
  221. **Tools:** Microsoft Threat Modeling Tool, OWASP Threat Dragon, draw.io
  222. ---
  223. ## Directory Structure
  224. ```
  225. ai-security-arsenal/
  226. ├── README.md
  227. └── skills/
  228. ├── android-pentest/ # Mobile - Android security
  229. ├── ios-pentest/ # Mobile - iOS security
  230. ├── dast-automation/ # Web - Dynamic testing
  231. ├── api-security/ # Web - API security
  232. ├── cloud-security/ # Cloud - Multi-cloud assessment
  233. ├── iac-security/ # Cloud - Infrastructure as Code
  234. ├── container-security/ # Cloud - Docker/Kubernetes
  235. ├── network-pentest/ # Network - Internal/AD testing
  236. ├── sast-orchestration/ # Code - Static analysis
  237. ├── sca-security/ # Code - Dependency scanning
  238. ├── llm-security/ # AI/ML - LLM security
  239. └── threat-modeling/ # Process - Threat analysis
  240. ```
  241. Each skill directory follows the same internal layout — see [Skill Architecture](#skill-architecture).
  242. ## Quick Start
  243. ### Mobile Testing
  244. ```bash
  245. # Android - Connect rooted device and start Frida
  246. adb devices
  247. adb shell "su -c '/data/local/tmp/frida-server -D &'"
  248. # "pentest com.example.app"
  249. # iOS - Connect jailbroken device
  250. iproxy 2222 22
  251. ssh -p 2222 root@localhost
  252. # "pentest com.example.app"
  253. ```
  254. ### Web Testing
  255. ```bash
  256. # DAST with Playwright MCP
  257. # "scan https://target.com for vulnerabilities"
  258. # API Security
  259. # "test API security for https://api.target.com"
  260. ```
  261. ### Cloud/Infrastructure
  262. ```bash
  263. # Cloud Security
  264. aws configure # or gcloud auth login
  265. # "assess AWS security for account 123456789"
  266. # IaC Security
  267. # "scan Terraform in ./infrastructure for security issues"
  268. # Container Security
  269. # "scan container image nginx:latest"
  270. ```
  271. ### Code Analysis
  272. ```bash
  273. # SAST
  274. # "scan this codebase for vulnerabilities"
  275. # SCA
  276. # "check dependencies for vulnerabilities"
  277. ```
  278. ## Requirements
  279. | Category | Skill | Key Requirements |
  280. |----------|-------|------------------|
  281. | Mobile | android-pentest | ADB, Frida, objection, rooted device, [Mobile MCP](https://github.com/mobile-next/mobile-mcp) (optional) |
  282. | Mobile | ios-pentest | Frida, objection, jailbroken device, iproxy, [Mobile MCP](https://github.com/mobile-next/mobile-mcp) (optional) |
  283. | Web | dast-automation | Playwright MCP, Nuclei, Python 3.8+ |
  284. | Web | api-security | Burp Suite, Postman, jwt_tool |
  285. | Cloud | cloud-security | Cloud CLI tools (aws, gcloud, az), ScoutSuite |
  286. | Cloud | iac-security | Checkov, tfsec, Terrascan |
  287. | Cloud | container-security | Trivy, Kubescape, kubectl |
  288. | Network | network-pentest | Nmap, BloodHound, Impacket, CrackMapExec |
  289. | Code | sast-orchestration | Semgrep, Bandit, CodeQL CLI |
  290. | Code | sca-security | Syft, Grype, language package managers |
  291. | AI/ML | llm-security | Python 3.8+, target LLM API access |
  292. | Process | threat-modeling | Draw.io or Threat Dragon (optional) |
  293. ## Usage with Coding Agents
  294. These skills are designed to work with AI coding agents that support skill/instruction systems. Invoke via natural language:
  295. ```
  296. You: pentest the banking app com.megabank.mobile
  297. Agent: [Executes android-pentest - spawns with Frida, bypasses SSL, extracts data...]
  298. You: scan our AWS account for security misconfigurations
  299. Agent: [Executes cloud-security - runs ScoutSuite, analyzes IAM, checks S3...]
  300. You: check this codebase for vulnerabilities
  301. Agent: [Executes sast-orchestration - runs Semgrep, Bandit, aggregates findings...]
  302. You: threat model the payment processing system
  303. Agent: [Executes threat-modeling - creates DFD, applies STRIDE, generates report...]
  304. ```
  305. **Structured output:** Each skill emits findings that validate against its `schemas/finding.json` — pipe directly into SARIF converters, issue trackers, or custom triage pipelines without reformatting.
  306. **Parallel execution:** Multi-target requests (multiple domains, multiple images, multiple clouds) are handled by spawning sub-agents per target, as documented in each skill's Sub-Agent Delegation section.
  307. **Compatible with:** Claude Code, Codex CLI, OpenCode, Aider, and other agents supporting markdown-based skill definitions.
  308. ## License
  309. For authorized security testing, penetration testing engagements, CTF competitions, and educational purposes only. Always obtain proper authorization before testing.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.