awesome-claude-security

awesome-claude-security

A Claude Code plugin marketplace for the full cybersecurity & GenAI-security lifecycle — from recon and threat modeling to detection engineering, GRC, and CISO-level strategy.

Docs site · Quick install · Getting started · What's inside · Taxonomy · Roadmap · Recipes · Contributing

A pentester knows which OWASP test bends a broken-access-control endpoint. An analyst knows which Sigma rule catches Kerberoasting. A red-teamer knows which payload coaxes an LLM past its guardrails, and a CISO knows how to turn that finding into board-ready risk. Your Claude Code doesn't — until you install the plugins that teach it.

Everything installs à la carte — one repo, but you install only the plugins you want, never "the whole thing." Want only LLM red-teaming? Install llm-security. Only threat modeling? Install threat-modeling.

Prefer a ready-made stack? Install a bundle and it auto-pulls its parts: a role bundle like pentester, or a domain suite like genai-suite. Granular and bundled both come from the same catalog — see docs/BUNDLES.md.

Stable: v1.1.0. The marketplace, taxonomy, templates, and a full first wave of 45 plugins (110 skills) are shipped and installable — see the roadmap for what's next. This is a community project; it is not affiliated with or endorsed by Anthropic. Contributions welcome — see CONTRIBUTING.

Quick install

In any Claude Code session:

/plugin marketplace add jassics/awesome-claude-security
/plugin install llm-security@awesome-claude-security

Then invoke a skill, e.g. /llm-security:owasp-llm-top10, or just describe your task and let Claude pick the right skill/agent. Full instructions: docs/INSTALL.md.

New here? docs/RECIPES.md shows end-to-end workflows — web pentest, incident response, vuln triage, design review, securing a GenAI feature — each chaining the right skills (and role bundles ship a one-shot command, e.g. /pentester:engagement).

What's inside

Plugins are grouped into four buckets (see the full taxonomy):

Bucket	What it is	Examples
Core	Cross-cutting capabilities every security task reuses	security-diagramming, security-reporting, integrations (Jira/Confluence/Drive)
Domain	Deep skillsets per security discipline	threat-modeling, web/mobile/cloud/k8s/network/infra security, SAST-SCA, OSINT, DFIR, detection engineering
GenAI security	Protecting AI/LLM systems from attackers	llm-security, RAG security, agentic-AI security, multimodal security
AI safety	Preventing AI systems from causing harm (a distinct discipline — see why)	ai-safety, ai-safety-engineer
Role	Persona bundles that combine domains + workflow	pentester, ai-safety-engineer, security analyst, engineer, architect, GRC, blue team, SOC/SIEM, CISO/CTO

Shipped today (45 plugins, 110 skills)

Core

1. security-diagramming (attack trees, DFDs, architecture diagrams, mindmaps, infographics)
2. security-reporting (findings, pentest reports, exec summaries, CVSS)
3. security-integrations (publish to Jira/Confluence/Drive)
4. security-knowledge (ATT&CK / OWASP / framework reference packs).

Domain

1. threat-modeling (STRIDE/PASTA)
2. web-app-security (OWASP Web Top 10, access control, injection)
3. api-security (OWASP API Top 10, BOLA/BFLA)
4. mobile-security (MASVS/MASTG)
5. sast-sca (static analysis + dependency/SBOM)
6. network-security (network pentest, segmentation, protocols)
7. osint (footprinting, exposure discovery, recon)
8. cloud-security (AWS/Azure/GCP posture, IAM, misconfig)
9. k8s-security (CIS/4Cs, RBAC, pod hardening)
10. infrastructure-security (IaC review, host hardening, secrets)
11. detection-engineering (Sigma/YARA, ATT&CK coverage, threat hunting)
12. dfir (incident response, forensic triage, IOCs)
13. threat-intelligence (CTI lifecycle, IOC enrichment, actor profiling)
14. vulnerability-management (triage, CVSS/EPSS/KEV prioritization, remediation SLAs)
15. supply-chain-security (dependency trust, SLSA/Sigstore provenance, CI/CD integrity)
16. claude-config-security (audit the Claude Code config itself — hooks/MCP/permissions/skills — via the agentscanner CLI).
17. secure-coding (Python/React outdated-function + secret/credential enforcement, git-hook blocking)

GenAI security

1. llm-security (OWASP LLM Top 10, prompt injection)
2. rag-security (retrieval poisoning, isolation)
3. agentic-ai-security (tool-permission audit, autonomy boundaries, MCP/agent-harness/A2A security)
4. multimodal-security (cross-modal injection)
5. mlops-security (ML supply chain, pipeline security, model serving).

AI safety (≠ security — see why)

1. ai-safety (harm modeling, safety evals, responsible red-team, bias/fairness, guardrails, RAI governance).

Roles (auto-install their stack)

1. pentester
2. red-team (adversary emulation, ATT&CK)
3. blue-team (threat-informed defense + purple teaming)
4. soc-siem (alert triage, monitoring)
5. security-analyst (investigation & analysis, T2/T3)
6. security-architect (secure-by-design, design review)
7. security-engineer (DevSecOps, harden, secure pipelines)
8. ai-safety-engineer
9. responsible-ai-officer (AI governance, EU AI Act risk-tiering).
10. grc (governance, risk & compliance — framework gap assessments, control mapping)
11. developer (secure-by-default coding companion, PRD/prompt security injection, pre-commit gate).

Executive (strategic tier)

1. ciso-toolkit (security strategy, cyber-risk quantification, board decks)
2. cto-security (secure-by-design at scale, tech-risk assessment).

Suites (one-shot bundles)

1. genai-suite (all GenAI security)
2. appsec-suite (web+api+mobile+SAST/SCA)
3. cloud-suite (cloud+k8s+infrastructure)
4. blueops-suite (detection+dfir+threat-intel)
5. ai-safety-suite (safety + GenAI security together).

The roadmap is fully shipped — see it for how the buckets fit together and where the project goes next.

Security vs. safety: the GenAI plugins protect AI systems from attackers; ai-safety prevents AI systems from causing harm even absent an attacker. They're complementary — most AI features need both.

Built-in capabilities

Several features are designed to be shared across every plugin:

• Diagrams & visuals — Excalidraw diagrams, mindmaps, attack trees, and infographics via security-diagramming.
• Reporting — repeatable report generation via security-reporting, from a single finding to a board deck.
• Integrations — push findings, reports, and diagrams to Jira, Confluence, and Google Drive (roadmap: security-integrations, leveraging the matching MCP servers).

Repository layout

.
├── .claude-plugin/marketplace.json   # the marketplace catalog
├── plugins/<plugin-name>/            # one installable plugin per directory
│   ├── .claude-plugin/plugin.json
│   ├── skills/<skill>/SKILL.md
│   ├── agents/<agent>.md
│   └── README.md
├── templates/                        # copy-to-start scaffolds for new plugins
└── docs/                             # taxonomy, roadmap, install, authoring guides

Contributing

This is meant to be a community resource. New domains, roles, skills, and agents are very welcome — start from templates/ and read CONTRIBUTING.md and docs/AUTHORING.md.

Scope & ethics

These assets are for authorized security testing, defensive security, detection, GRC, education, and CTF use. Skills assume you have permission to test the systems involved. See CONTRIBUTING.md.

License

GPL-3.0.