AegisMind

Forge your security mind. A living knowledge base for cybersecurity, AI security, offensive tooling, agent stacks, and Anthropic-style security skills — curated in the open.

Author: Ayush Kumar · DevRel · AI Security · NeuralVerse · RuntimeWall

---

What is this?

AegisMind (aegis = shield) is my personal security learning vault — not a course, not a dump of links. It is a structured, version-controlled notebook where I collect:

• Offensive & defensive cybersecurity notes
• AI / LLM security research & red-teaming playbooks
• AI tooling (agents, MCP, RAG, eval frameworks)
• AI security tooling (Promptfoo, guardrails, runtime governance)
• Anthropic cybersecurity skills & agent skill patterns
• Resources, cheatsheets, and battle-tested workflows

Everything here is learned in public. Pull requests and issues welcome.

---

Quick navigation

Section	What's inside	Start here
Cybersecurity	Recon, web vulns, cloud, bug bounty, red team	cybersecurity/README
AI Security	LLM red teaming, prompt injection, ASI framework	ai-security/README
AI Tools	LLMs, agents, MCP, RAG, local stacks	ai-tools/README
AI Security Tools	Promptfoo, Deepteam, guardrails, eval harnesses	ai-security-tools/README
Anthropic Skills	Agent skills, security skill patterns, MCP workflows	anthropic-skills/README
Resources	Books, courses, reports, wordlists, templates	resources/README

---

Repository map

aegis-mind/
├── README.md                 ← you are here
├── docs/
│   ├── getting-started.md    ← how to use & contribute
│   └── CONTRIBUTING.md
├── cybersecurity/            ← classic security learning
├── ai-security/              ← LLM / agent security
├── ai-tools/                 ← builders' toolkit
├── ai-security-tools/        ← security-specific AI tooling
├── anthropic-skills/         ← skills, agents, MCP patterns
└── resources/                ← external references & cheatsheets

---

Learning tracks

Pick a path — or mix them. Each folder has its own index.

Track 1 — Cybersecurity foundations

Recon → Web/API → Cloud → Bug Bounty → Reporting

→ cybersecurity/

Track 2 — AI security

Threat model → Red team LLMs → Guardrails → Runtime governance

→ ai-security/

Track 3 — Build + secure agents

Local LLMs → Agents & MCP → Eval with Promptfoo → Ship with RuntimeWall

→ ai-tools/ + ai-security-tools/

Track 4 — Anthropic-style skills

Skill anatomy → Security skills → MCP tool design → Agent orchestration

→ anthropic-skills/

---

Featured stacks

Domain	Tools & frameworks
Offensive	Burp, ffuf, nuclei, custom wordlists
LLM eval	Promptfoo, Deepteam, custom probes
Agents	MCP servers, Cursor/Codex skills, agent runtimes
Runtime	Sandboxing, session replay, tool allowlists
Writing	NeuralVerse blog patterns, report templates

---

How content is organized

Every note follows a simple template so future-you (and readers) can scan fast:

1. Goal — what problem this solves
2. Prerequisites — what you need before starting
3. Steps / commands — reproducible, copy-paste friendly
4. Validation — how to know it worked
5. References — papers, blogs, disclosed reports
6. My notes — gotchas, failures, real engagement lessons

See docs/getting-started.md for the full template.

---

Status

Area	Progress
Cybersecurity	🟡 Growing
AI Security	🟡 Growing
AI Tools	🟡 Growing
AI Security Tools	🟡 Growing
Anthropic Skills	🟢 Active
Resources	🟡 Growing

Legend: 🟢 Active · 🟡 Growing · ⬜ Planned

---

Connect

Link
GitHub	@Ayush7614
Blog	NeuralVerse
Portfolio	ayushbuilds.dev
LinkedIn	ayush-kumar
Mentorship	Topmate

---

License

MIT — see LICENSE. Learning notes are shared freely; never commit secrets, live engagement credentials, or client data.

---

AegisMind · shield your stack · sharpen your mind · ship in the open
_{© 2026 Ayush Kumar · built with curiosity}