Full OAuth2-pattern REST API built from scratch with Python's standard library. JWT signing, PBKDF2 password hashing, role-based access control, brute-force protection, and token rotation with no frameworks or pip installs.

CLI tool that catches leaked credentials in source code before they ship. 20+ detection rules, Shannon entropy analysis, git pre-commit hook, and SARIF output for GitHub Code Scanning. Python standard library only.

Tests Retrieval-Augmented Generation pipelines for indirect prompt injection, context poisoning, and data exfiltration. Includes a built-in TF-IDF retrieval engine so you can test without any external vector database.

Red-teams LLM system prompts by firing 40 injection payloads across 10 attack categories and scoring resistance. Canary-based detection with letter grades and weighted risk scoring for CI pipeline integration.

Scans LLM responses for PII leakage, hallucinated claims, policy violations, and toxic content before they reach the end user. Uses the LLM-as-judge pattern for hallucination detection with offline regex checks as a baseline.

Collection of automated incident response playbooks that handle initial triage for suspicious IPs, brute force attacks, IOC enrichment, and log anomaly detection.

Python command-line tool that analyzes .eml files for phishing indicators, extracts IOCs, checks email authentication, and scores risk.

A real-time threat intelligence dashboard that aggregates IP, domain, and hash indicators across security feeds into a single consolidated lookup interface.

An access-controlled security dashboard built with Next.js and Supabase Row-Level Security (RLS) to restrict administrative operations and safeguard sensitive records.

A zero-trust, client-side configuration analyzer that audits AWS JSON policies against 15 CIS Benchmark rules with line-by-line syntax flagging and instant remediation guidance.