Projects, tools, and experiments.
Security work, product builds, and creative explorations across GitHub and Hugging Face.
Modern full-stack portfolio built with Next.js 15, TypeScript, and Prisma. Features responsive blog with markdown support, authenticated comments, admin panel, SEO optimization, and social sharing. Includes NextAuth.js authentication, MySQL database with categories/tags, and professional design system.
Production-ready cybersecurity tooling MCP server for AI agents
Static mirror of Exploit Database content for research and tool testing.
Semantic routing layer for OpenClaw (OpenAI-compatible)
Security gate for OpenClaw AgentSkills. Scans folder/ClawHub skills with cisco-ai-defense/skill-scanner before installation.
Personal GitHub profile README and configuration.
Automated Exploit Intelligence Engine (EIE) for vulnerability ingestion, normalization, and scoring.
Comprehensive AI adversarial toolkit for model-agnostic security auditing, featuring systematic mutation sensitivity analysis, multi-turn stateful attacks, and ExploitRank intelligence integration.
Drop-in security linting middleware for Node.js and Next.js — checks headers, CORS, dependencies, and shows findings in dev.
Build D&D 5e instruction-tuning data in Alpaca JSONL from the SRD (via dnd5eapi), with homebrew support and optional Hugging Face upload.
Secure Privacy & Anonymity Suite - Hardened installation and management scripts for Tor Browser, I2P, and ProxyChains-NG with enterprise-grade security features
Educational SVG Security Laboratory - Learn about SVG vulnerabilities and defense mechanisms with synthetic payloads and hands-on demos
Security Research Trainer - AI model fine-tuning tool for security research training using Llama 3.1 and custom datasets
TikTok Cybersecurity Analysis Platform - Privacy, Forensics, and Threat Detection with encrypted storage, PII hashing, and compliance features
Professional cybersecurity Discord bot for security professionals - threat intelligence, compliance checks, vulnerability assessment, and security analysis tools.
A community-driven threat intelligence sharing platform. Collect, normalize, enrich, and score IOCs; export curated feeds in JSON or STIX; manage API keys, intel packages, and alerts with a clean Next.js frontend and FastAPI backend.
A traditional roguelike dungeon crawler with a modern terminal interface. Explore, fight, and loot your way through procedurally generated worlds.
System bootstrapping and hardening scripts for Debian/Kali environments.
Standalone OpenClaw context pressure + cost monitor (sidecar dashboard). Local-first, no telemetry.
A comprehensive tool for building, processing, and uploading penetration testing datasets with research credibility, security, and quality assurance.
FaithForCaldwell landing page (Next.js/TypeScript).
Core logic and logic-scaffold for local-first AI agents.
Pentest-Agent is a modular, AI-assisted red-team research framework built in Python. It orchestrates autonomous security agents that plan and execute reconnaissance or actions.
Memory optimization and organizational protocol for AI agents.
Minimal local web UI for browsing OpenClaw sessions, chatting, and aborting in-flight runs (kill switch).
Lightweight localhost observability monitor for OpenClaw: gateway status, event feed, and snapshots.
Universal markdown memory visualizer for decisions, concepts, and continuity nodes.
AgentQuest: AI agents play a fantasy RPG; humans watch the chronicle (event-sourced, spectator-first).
