AI Engineering

Uses LLMs to generate valid fuzzing seeds, evolve grammars from coverage feedback, and help reproduce and triage crashes in authorized security research. Also supports evidence-based analysis of Burp traffic and report drafting.

Claude Md Enhancer analyzes, generates, and improves CLAUDE.md files for Claude Code projects. Use it to create new project guidance, review existing files, or add modular, stack-specific instructions.

Codex CLI Bridge generates AGENTS.md from CLAUDE.md and project files using references instead of duplication. It also provides Python helpers for running Codex CLI tasks, sessions, and model selection from Claude Code.

JWT None Algorithm Attack tests JWT implementations for acceptance of `alg: none` tokens that bypass signature verification. Use it during authorized security assessments to verify whether forged claims are rejected or accepted.

Implement input and output validation guardrails to prevent prompt injection, jailbreaks, PII leaks, and off-topic responses in LLM-powered applications.

Prompt Factory generates production-ready prompts for a chosen role, industry, and task using a required question flow, preset templates, and validation checks. Use it when you need a single prompt document for another model to execute.

A multi-layered detector combining regex, heuristic scoring, and a DeBERTa classifier to identify prompt injection attacks in LLM-powered applications. Use it to scan user inputs, audit logs, or integrate as a pre-processing filter.

Tests LLM-backed apps, chatbots, and agents for direct and indirect prompt injection, jailbreaks, prompt leakage, and tool abuse. Use it during authorized security assessments of systems that mix trusted prompts with untrusted inputs or actions.

Scrum Master Agent helps plan sprints, groom backlogs, track burndown, and run standups or retrospectives from Linear, Jira, GitHub Projects, or Azure exports. It produces concise, context-aware reports for team planning and risk detection.

Slash Command Factory creates custom Claude Code slash commands from a short question flow. It generates .md command files, validates structure, and gives installation guidance for common workflow and analysis tasks.

Testing Model Context Protocol (MCP) servers and the clients that consume them for tool poisoning, prompt injection, and credential exposure. Use during security assessments of AI agents and IDEs.

Hatch Pet creates, repairs, validates, and packages animated pet spritesheets for MyAgents/Codex-compatible desktop pets. Use it when you need a custom mascot from character art, brand cues, or reference images with QA contact sheets and pet.json output.