Governance & Ecosystem Audit
We audit your ecosystem as a graph
The questions a cap-table can't answer — which holder clusters could collude to capture governance, where an incentive loop feeds back on itself, what the shortest path is from an admin key to protocol funds — run as graph traversals over your real ecosystem.
A cap-table lists who holds what; it can't tell you who could collude to control the vote. We model your ecosystem as a graph — holders, contracts, treasuries, incentive loops and governance paths as nodes and edges — and run the traversals and graph algorithms that surface capture risk, sybil clusters and admin-key exposure a flat table hides.
Quorum looks safe on paper
No single holder crosses the threshold — but three of them together do, and they're already connected. A percentage column can't see that.
Sybil clusters read as many holders
Dozens of addresses funded from one source count as decentralisation in a table and as one voter in reality.
Admin keys quietly reach the funds
An upgrade role two hops from the treasury is a governance risk, not just a technical one — and it's a path, not a number.
Dilution is assumed to fix it
"Emissions will decentralise us over time" is a claim nobody simulates. Often the insider bloc still clears quorum a year later.
What we build
Collusion & capture analysis
Which holder clusters could combine to cross a governance threshold — reachability plus community detection (GDS Louvain / WCC) over the real holder graph.
Whale & quorum stress-test
Whale-concentration and quorum analysis — including whether insiders still clear quorum after a year of unlocks — modelled with the graph⇄simulation bridge.
Incentive-loop & admin-key paths
Where an incentive loop feeds back into itself, and the shortest path from an admin key or upgrade role to protocol funds.
Capture-over-time simulation
radCAD Monte-Carlo seeded from the live graph and written back as a scenario subgraph — a structural question asked about simulation output.
What you walk away with
Deliverables, not a slide deck.
- A graph model of the live ecosystem — holders, contracts, roles, treasuries
- Collusion-reachability and community analysis (GDS Louvain / WCC) over the real holder graph
- Whale-concentration and quorum stress-test, including capture-over-time
- Incentive-loop and admin-key-to-funds path analysis
- A ranked capture report — with the exact Cypher query behind every finding, so you can re-run it
How we engage
Map
We reconstruct the ecosystem — holders, contracts, roles — as a graph.
Traverse & simulate
GDS algorithms for structure, radCAD for capture-over-time.
Capture report
The risks, ranked, with the exact query behind each finding.
Where it matters most
The situations this is built for.
Before handing over governance
You're about to transfer control to a DAO and need to know whether that control is genuinely distributed.
After a contentious vote
A proposal passed and it doesn't feel organic. We reconstruct who actually decided it, and how they're connected.
Unlock-cliff planning
A large tranche is about to vest. We model what it does to voting power before it lands, not after.
Proof
On $GUILD, the graph⇄simulation bridge answered a question a flow model can't: after a year of vesting, does the insider bloc still clear quorum? Insider share of float fell from 86% to 57% — still a majority, still far above the 5% quorum.
Questions
Model it. Prove it. Then launch.
Tell us what you're building. We'll come back with a tailored proposal.
Book a technical call