# Christopher Hammer AI Portfolio - Mini Demo Guide

This guide is for reviewers who want to verify functional systems quickly.

## 1. HammerLock AI

Live system: https://hammerlock.vercel.app  
Repo: https://github.com/christopherlhammer11-ai/hammerlock
Local desktop artifact: `/Users/miahammer/Documents/hammerlock/dist-electron/HammerLock-AI.dmg`

What to check:
- Product positioning for privacy-first local AI.
- Agent architecture, vault, encryption, and offline-first story.
- Repo tests/build and README for implementation direction.
- Desktop packaging path exists; public release still needs signed/notarized hosting.

Why it matters:
HammerLock shows product judgment around the core trust problem in AI assistants: useful automation without surrendering sensitive memory and workflows to a remote black box.

## 2. Generic CRM

Live system: https://generic-crm.vercel.app  
Repo: https://github.com/christopherlhammer11-ai/generic-crm

What to check:
- Contact, pipeline, and dashboard workflow.
- Local AI assistant direction with Ollama.
- Next.js product structure and deployability.

Why it matters:
Generic CRM shows that the AI work can sit inside a real business workflow instead of living only as libraries or experiments.

## 3. Genesis Node API

Health endpoint: https://genesis-node-api.vercel.app  
Marketplace UI: https://genesis-marketplace.vercel.app  
Repo: https://github.com/christopherlhammer11-ai/genesis-node-api

What to check:
- Open the health endpoint and confirm JSON service metadata.
- Run discovery against the live API:

```bash
curl -L -sS -X POST https://genesis-node-api.vercel.app/v1/discover \
  -H 'content-type: application/json' \
  -d '{"query":"guardian"}'
```

Expected result:
The API returns the Tool Use Guardian skill with package URL, pricing, interface metadata, and performance fields.

Why it matters:
Genesis Node API shows agent-to-agent capability discovery: agents finding installable skills through a service boundary rather than a hardcoded script.

## 4. Tool Use Guardian

Live browser demo: https://2026-04-21-that-s-a-full-green-run.vercel.app/tools-demo.html  
Repo: https://github.com/christopherlhammer11-ai/tool-use-guardian

What to check:
- Test suite for timeouts, retries, JSON repair, and structured failures.
- README examples for wrapping unreliable tools.

Why it matters:
Tool Use Guardian shows understanding of a real agent failure mode: tool calls are where demos become production systems or break under pressure.

## 5. Prompt Condenser

Live browser demo: https://2026-04-21-that-s-a-full-green-run.vercel.app/tools-demo.html  
Repo: https://github.com/christopherlhammer11-ai/prompt-condenser

What to check:
- CLI and library shape.
- Preservation of code, JSON, URLs, and user intent while reducing prompt size.

Why it matters:
Prompt Condenser shows awareness that token volume is cost, latency, and reliability surface.

## 6. Go-to-Market Plan

Plan: https://2026-04-21-that-s-a-full-green-run.vercel.app/go-to-market.md
Finish report: https://2026-04-21-that-s-a-full-green-run.vercel.app/project-finish-report.md

What to check:
- Buyer segmentation for founders, lawyers, and financial advisors.
- Pricing logic for Free, Pro, and Team.
- Launch sequence for Hacker News, Product Hunt, LinkedIn, and direct outreach.
- Demo-ready project status and verification counts.

Why it matters:
The product story is now connected to an actual market: professionals who need AI
help but cannot casually expose sensitive work to cloud chat histories.

## Shared Terms

- AI agent: A model-driven system that decides next actions and may call tools, read context, update memory, or ask for human input.
- Tool calling: A model requesting a structured external action such as API calls, search, file work, browser automation, or database work.
- Verification: Checking outputs before trusting them, including JSON validation, URL checks, claim checks, and confidence scoring.
- Local-first: Product architecture where important state and workflows stay on the user's device by default.
- Guardrail: A rule, validator, timeout, permission boundary, or review step that prevents unsafe or invalid action.