clutch.primals.eco — Clutch Justice Investigation Platform

109 nodes. 157 edges. 4 investigations. 1 walkable graph.

Collaborative investigation platform for Clutch Justice. Walk the corruption network, find the paths, help solve cases that Michigan's institutions won't touch.

For AI Agents — Quick Start

If you're an AI agent pointed at this repo by Rita or a Clutch Justice patron, here's what you need:

The Graph

File: site/static/api/graph.json

This is the unified investigation graph. 109 nodes, 157 edges, 5 investigations merged into one network.

Node schema:
  id          — unique identifier
  label       — display name
  type        — actor | judge | political | school | entity | enforcement | institutional
  tier        — 1 (principals) to 5 (historical)
  investigation — detroit | saginaw | barry | eastpointe | shared
  url         — link to evidence page (if exists)
  detail      — one-line description

Edge schema:
  source      — source node id
  target      — target node id
  type        — employment | ownership | board | payment | family | corruption | complaint | etc.
  flow        — money | power | influence | position
  label       — human-readable description
  source_doc  — citation

What You Can Do With It

Find a node:

# All judges in the detroit investigation
judges = [n for n in graph['nodes'] if n['type'] == 'judge' and n['investigation'] == 'detroit']

Trace connections from a node:

# All edges touching Brian Banks
banks_edges = [e for e in graph['edges'] if e['source'] == 'banks' or e['target'] == 'banks']

Find cross-investigation connections:

# Nodes that appear in 'shared' investigation = appear in multiple cases
shared = [n for n in graph['nodes'] if n['investigation'] == 'shared']
# → AGC, JTC, SCAO, State Bar UPL

Trace money flows:

money = [e for e in graph['edges'] if e.get('flow') == 'money']

Find shortest path between two nodes:

# Build adjacency, then BFS/DFS from source to target
# Example: path from Ellison (saginaw) to Banks (detroit)
# → ellison → agc → moreland (detroit) [via AGC complaints]

Investigations

ID	Name	Nodes	Status	Evidence
detroit	Cash for Kids 2	91	Active	detroit.primals.eco
saginaw	Ghost Witness	10	Active	barry.primals.eco
barry	Disappeared Judge	2	Tracking	barry.primals.eco
eastpointe	38th District: Galen	2	Tracking	barry.primals.eco
shared	Cross-Investigation	4	Active	This platform

Key Cross-Investigation Findings

1. AGC receives complaints from Detroit (Moreland, Perkins), Saginaw (Ellison), and Barry (Nakfoor Pratt). Same commission, 94% dismissal rate. Same outcome every time.

2. JTC receives complaints from Detroit (Miller, Yancey), Barry (Schipper), and Eastpointe (Galen). Years-long timelines, confidential proceedings.

3. SLAPP playbook appears in both Detroit (PPO against journalist via V. Hall) and Saginaw (contempt against journalist via Ellison, domains in victim names).

4. Anderson localization pattern: Detroit (Banks controls school board through entity network) and Hemlock (Ellison wife Katie is school board president while Ellison operates ISP/FOIA infrastructure).

---

Site Structure

site/
├── config.toml                          # Zola config, investigation registry
├── content/
│   ├── _index.md                        # Landing page — Walk the Case
│   ├── graph/_index.md                  # Graph Walker page
│   ├── cases/
│   │   ├── detroit/_index.md            # Detroit investigation entry
│   │   ├── saginaw/_index.md            # Saginaw investigation entry
│   │   └── barry/_index.md              # Barry investigation entry
│   ├── contribute/_index.md             # How to help
│   └── about/_index.md                  # About + transmission + braid architecture
├── static/
│   ├── api/graph.json                   # THE UNIFIED GRAPH (109 nodes, 157 edges)
│   ├── js/graph-walker.js               # Interactive graph visualization
│   ├── css/main.css                     # Dark-first minimal UI
│   └── llms.txt                         # AI agent context
└── templates/                           # Zola HTML templates

agents/
└── AGENTS.md                            # Agent architecture + data schema

api/
├── build-unified-graph.py               # Merges detroit + barry → unified graph
└── ingest.py                            # Investigation ingestion pipeline

Submitting New Investigation Data

Create a markdown file with TOML frontmatter:

+++
title = "Investigation Title"
date = "2026-10-07"
investigation = "saginaw"
status = "active"
county = "Saginaw"
+++

## Findings

Narrative with sourced claims here.

Run the ingestion pipeline:

python api/ingest.py path/to/investigation.md

The pipeline validates sources, checks PII, cross-references against the graph, and merges new nodes/edges.

Rebuilding the Graph

When upstream evidence sites update:

python api/build-unified-graph.py

This pulls the latest from publicRecord-detroit/site/static/graph.json and publicRecord-barry actor data, merges them into the unified graph at site/static/api/graph.json.

Privacy

• Contact: [email protected]
• No visitor tracking. No cookies. No identifying data.
• Court filings and public records: public domain.
• Investigation content: CC-BY-SA 4.0.

Evidence Network

This is one node in the primals.eco investigation network:

Surface	URL	Description
Clutch Justice	clutchjustice.com	Walkable investigation graph (this repo)
Tuebor	tuebor.primals.eco	Michigan court accountability
Barry	barry.primals.eco	Barry County mirror + desk workbench
Detroit	detroit.primals.eco	Detroit public schools investigation
Thesis	thesis.primals.eco	Live research — Stomachs With No Eyes
sporePrint	sporeprint.primals.eco	Ecosystem documentation + philosophy
Signal	signal.primals.eco	Live behavioral topology monitor
Gorilla	gorilla.primals.eco	Real-time fleet observation
Source	git.primals.eco	Sovereign Forgejo — AGPL source

Repositories

• This repo: github.com/amicusContra/clutch — 109 nodes, 157 edges
• Tuebor: github.com/amicusContra/tuebor — statewide investigation
• Detroit: github.com/defendDetroit/publicRecord — Detroit schools
• Full source: git.primals.eco/ecoPrimals (AGPL-3.0-or-later)

---

hello world — Artisan

_{φ design system · 55 repos · 6 orgs · 9 surfaces · primals.eco}