Strategy Hub
Firma Network
Build Roadmap v2 — Moral Framework Foundation
February 2026 · Internal
Firma Network — Complete Build Sequence

Agents of
Generosity,
Not Extraction

Every person in the Firma ecosystem has an agent. Every agent operates from a declared moral framework. You cannot participate — cannot hold reputation, deploy a network state, contribute compute, or receive grants — without first declaring the values your agent acts from. The moral framework is open source. The community defines it. The network enforces it. This is not a neutral infrastructure layer. It is a network built on the premise that how you build matters as much as what you build.

Layer 0
0
Weeks 1–4 · Runs parallel to everything

The Moral Framework Standard

This is the entry condition for the entire network. Before you can hold a Firma identity, before your agent can act, before you can receive a grant or deploy a network state or contribute compute — you declare which moral framework you're operating under. The framework is open source. Firma doesn't own what "good" means. The community defines it, versions it, and evolves it. But participation requires declaration. An agent without a declared framework cannot participate. This is what makes the network an intentional ecosystem rather than neutral infrastructure.

Step 0.1
Draft the Firma Moral Framework Standard v1.0
Founders Open Source NSS Input
Write the first version of the open-source moral framework that every agent must declare before participating. This is not a terms of service. It is a values declaration. It covers: orientation toward generosity over extraction, commitment to building for collective betterment, transparency of intent, non-exploitative relationships with the people and ecosystems you interact with. v1.0 is drafted by Firma's founders with input from NSS — Balaji's community has spent more time thinking about network state ethics than anyone. The framework is published as an open-source document on GitHub with a versioning protocol so the community can evolve it through governance.
Why open source matters here: If Firma owns the moral framework, the network is only as trustworthy as Firma. If the community owns it, the network is trustworthy because the values belong to everyone inside it. Open source is not a technical choice here. It is a governance choice. The framework's legitimacy comes from the fact that no single entity controls it.
Framework v1.0 published on GitHub Versioning protocol established Community governance process defined
Step 0.2
Define the three framework dimensions every agent declares
Founders LightCI Engineering
The declaration isn't binary — it's three dimensions that define how your agent behaves across the ecosystem. Orientation — generosity vs extraction spectrum, where you declare your intent (giving, building, trading, governing). Scope — local (my network state), regional (my union), global (all of humanity). Constraint — what your agent will never do regardless of instruction: extract from vulnerable actors, deceive, accumulate without contributing back. These three dimensions are stored on-chain as part of your soulbound identity. Your agent's behavior is evaluated against them. Deviation from declared framework affects reputation score.
The enforcement mechanism is reputation, not punishment. If your agent acts against your declared framework, your reputation score reflects the divergence. Other agents and institutions can see the gap between what you declared and how you acted. The network doesn't ban you — it just becomes transparent about who you are. That transparency is the enforcement.
Three-dimension declaration schema On-chain declaration contract spec Reputation deviation tracking spec
Step 0.3
Publish framework as entry condition — no declaration, no participation
Policy On-Chain LightCI
The framework declaration is technically enforced at the identity layer — LightCI builds the declaration step into the identity onboarding flow. You cannot complete identity creation without completing your framework declaration. The declaration is signed on-chain. Your agent is initialized with your declared framework as its operating parameters. Undeclared wallets cannot receive grants, cannot participate in governance, cannot deploy to Firmamint infrastructure, cannot access Seedbase matching. The gate is the declaration. Everything inside the gate operates from declared values.
Declaration required in identity flow Undeclared wallets gated from participation Framework stored in identity contract
I
Generosity as Default Orientation
Every agent in the network is initialized with a bias toward giving, building, and contributing — not toward accumulation, extraction, or competitive displacement. This doesn't mean agents can't trade or earn. It means the frame through which they operate is "what does this create for the collective" before "what does this return to me."
II
Declared Constraints Are Inviolable
When you declare what your agent will never do, that declaration is public and on-chain. Your reputation score tracks the gap between declaration and behavior. There is no private mode. There is no test environment where the framework doesn't apply. Every action your agent takes is evaluated against what you said you stood for.
III
The Framework Evolves Through Governance
Firma v1.0 is a starting point, not a fixed doctrine. The community can propose amendments, submit new versions, and adopt updated frameworks through NSA governance. When the framework updates, participants choose whether to re-declare under the new version or remain under their current declaration. The evolution is consensual and auditable.

The moral framework is not a filter on who can join. It is a declaration of how you will act while you are here. The network does not judge your values. It makes them visible.

Layer 1
1
Weeks 1–3 · Parallel to Layer 0

Legal Entity + Identity Primitive

Two things happen in parallel here that are both prerequisites for everything else. The NSA legal entity is incorporated — the governing wrapper for Firmamint, the signing entity for Mintware's EIR contract, the treasury holder. And LightCI begins building the soulbound identity primitive that every person in the ecosystem carries — one identity, one reputation, traveling across Seedbase, NSA governance, NSSim builds, and the reserve layer. The legal entity is the organizational foundation. The identity primitive is the personal foundation.

Step 1.1
Incorporate the Network State Alliance (NSA) legal entity
Legal Firma
Form the NSA as a Wyoming DAO LLC or Cayman foundation. The NSA is the governing entity for Firmamint — it holds the chain's multisig treasury, signs contracts with contributors, and issues economic rights. Every project that enters the collective — Mintware, SRI, NSS cohorts, Base builders, future EIRs — gets their rights issued by the NSA through this legal structure. Wyoming DAO LLC is faster and cheaper. Cayman foundation is better if you expect international sovereign partners to be signing counterparties. Decide based on where SRI and Balaji are most comfortable.
NSA entity filed Operating agreement drafted Multisig wallet structure defined Mintware EIR term sheet ready
Step 1.2
Commission LightCI — soulbound identity + reputation primitive
LightCI Identity Firmamint
LightCI receives the full identity and reputation brief (see separate document). Their deliverable is the single piece of infrastructure everything else depends on: one soulbound identity per person, reputation dimensions that accumulate across all contexts (Seedbase giving, NSA governance, NSSim builds, compute/reserve contributions), a privacy layer that lets people prove their reputation tier without exposing the full history, and the moral framework declaration built into onboarding. Full IP transfers into the Firma Collective trust on completion. LightCI earns economic rights in the collective for their contribution.
Why LightCI specifically: Identity and reputation infrastructure is their domain — verifiable credentials, ZK proofs for privacy-preserving reputation, cross-context portability. You don't want a generalist shop solving this. The hard problems are already solved in their toolkit.
LightCI contracted Identity brief delivered IP transfer terms in contract LightCI economic rights allocated
Step 1.3
Sign Mintware as first Entrepreneur in Residence
Mintware NSA Seedbase
Formal EIR agreement signed. Mintware builds Seedbase on Firmamint and earns economic rights in the Firma Collective. Their reputation in the ecosystem is seeded from day one — they are the first entry in the on-chain economic rights registry. Mintware starts speccing Seedbase against the Firmamint testnet while LightCI builds the identity layer — both are ready to integrate at the same time.
EIR agreement signed Seedbase spec started Mintware rights allocated on-chain
Layer 2
2
Weeks 4–12 · Built by senior AI hire + LightCI

Every Person Gets an Agent

The personal agent is not a chatbot. It is the on-chain expression of your identity and your declared moral framework. When you give through Seedbase, your agent acts. When you vote in NSA governance, your agent acts. When you contribute compute to the reserve layer, your agent settles the transaction. The agent is how your reputation accumulates — not by what you say but by what it does on your behalf, measured against what you declared you stood for. One agent per person. Tied to the soulbound identity LightCI builds. Operating from the moral framework you declared in Layer 0.

Personal Agent Architecture — What It Is and What It Does
IDENTITY ANCHOR ├── Soulbound token from LightCI identity layer — one per person, non-transferable ├── Moral framework declaration stored in identity contract — orientation, scope, constraints ├── Agent initialized with declared framework as operating parameters └── Agent cannot act outside declared framework without reputation consequence REPUTATION ACCUMULATION — same identity, all contexts ├── Seedbase — giving history, mission alignment, donor reliability, LP participation quality ├── NSA governance — voting record, proposal quality, participation consistency ├── NSSim builds — territories designed, SRI validation scores, projects deployed on-chain ├── Reserve layer — compute contributed, energy capacity, asset holdings └── All contexts write to the same reputation registry — one profile, all contexts visible WHAT THE AGENT ACTUALLY DOES ├── Executes giving and matching logic in Seedbase on your behalf ├── Participates in governance votes according to your declared preferences ├── Advises on NSSim territory design — your personal advisor, not a generic tool ├── Manages compute and reserve contributions — settlement, routing, fee collection └── Surfaces reputation signals to partners — "this person has given consistently for 18 months" FRAMEWORK ALIGNMENT SCORING ├── Every agent action is evaluated against the declared moral framework ├── Alignment score: how closely your agent's actions match what you declared ├── High alignment → reputation grows faster, access expands, LP matching improves ├── Divergence from framework → visible in reputation profile, reduces trust signals └── The gap between declaration and behavior IS the reputation signal PRIVACY LAYER — ZK credentials ├── Full profile on-chain and auditable by NSA governance ├── ZK credential: prove "I have given consistently for 12+ months" without revealing who to ├── ZK credential: prove "my alignment score is above threshold" without revealing full history └── People control what they share — but cannot hide what they've done from the network itself
Step 2.1
Hire senior AI engineer — owns agent architecture and NSA infrastructure agent
Senior AI Hire Agent Arch Firmamint
This is the most important engineering hire in the entire stack. The senior AI engineer owns two things: the personal agent architecture (working with LightCI's identity layer) and the NSA infrastructure agent (the agent that settles trades, routes fees, and manages treasury between network states). They set the standards that every other agent in the ecosystem — Mintware's Seedbase agent, the NSSim advisory agent, future EIR agents — conforms to. They are not a contractor. They are staff. The agent architecture is too foundational to live outside the organization.
What to look for: Deep experience with agentic frameworks (LangChain, LlamaIndex, or lower-level), on-chain agent patterns (ERC-4337 account abstraction is relevant here), and strong opinions about agent safety and alignment. The last one matters — you are building agents of generosity. The engineer building the framework needs to understand why that constraint is a feature, not a limitation.
Senior AI engineer hired Agent architecture spec written Integration spec with LightCI identity layer
Step 2.2
Build and deploy personal agent framework on Firmamint
Senior AI Hire LightCI Firmamint
Senior AI engineer and LightCI build together: the personal agent is initialized from the soulbound identity token, reads the on-chain moral framework declaration, and has write access to the reputation registry. The agent ships with four base capabilities at launch: giving/matching (Seedbase), governance participation (NSA), advisory (NSSim Sandbox mode), and contribution settlement (reserve layer). Each capability is modular — Mintware extends the giving module for Seedbase, NSS extends the advisory module for the governance designer, without touching the core architecture.
Personal agent contracts deployed on Firmamint Four base capabilities live Reputation registry writing correctly Framework alignment scoring active ZK credential layer integrated

Your agent is not a tool you use. It is the record of how you acted — measured against what you said you stood for. Reputation is not what you claim. It is what your agent did.

Layer 3
3
Weeks 3–8 · Parallel to Layers 0–2

Firmamint — The Chain
Goes Live

Firmamint is an Optimism L2. It does not need to be complete to be live. It needs: the chain running with FIRMA as gas, the Basemint bridge carrying USDC to and from Base, the NSA multisig as treasury, and the economic rights + reputation registry contracts. Everything built in Layers 0, 1, and 2 deploys here. Everything built in Layers 4, 5, and 6 deploys on top of it.

Step 3.1
Deploy Firmamint — OP Stack L2 with FIRMA gas token
Blockchain Eng OP Stack
One senior blockchain engineer who has deployed an OP Stack chain deploys Firmamint. Testnet in days, mainnet configuration complete within two weeks. FIRMA is the native gas token. NSA multisig is the chain's governing address. This engineer is the same person who deploys the Basemint bridge and the economic rights contracts — they own the chain layer end to end.
Firmamint testnet live FIRMA gas token deployed NSA multisig as chain admin
Step 3.2
Build Basemint bridge — native USDC two-way bridge to Base
Blockchain Eng Coinbase / Base
The Basemint bridge is the circulatory system. USDC flows in and out of Firmamint through it — grants deployed, fees settled, contributors on-ramped and off-ramped. Built on the Optimism Superchain standard bridge architecture — not custom, not a third-party bridge, not wrapped tokens. Native USDC across a named architectural bridge. This is also the infrastructure that Coinbase/Base is named as a partner in — the bridge being named "Basemint" is the partnership, not just a technical choice.
Basemint bridge deployed USDC flowing Firmamint ↔ Base Bridge audit scoped
Step 3.3
Deploy NSA treasury + economic rights + reputation registry contracts
Blockchain Eng LightCI Identity
Three contracts that make the collective function: NSA treasury (USDC + FIRMA in multisig), economic rights registry (who contributed what, what they earn), and reputation registry (what LightCI's identity layer writes to). Mintware's EIR rights are the first entry. LightCI's contribution is the second. Every person and project that joins after this has their contribution and their reputation recorded in the same immutable history. The chain's history is the collective's record of who built what.
NSA treasury live Economic rights registry deployed Reputation registry integrated with LightCI Firmamint mainnet live
Layer 4
4
Weeks 8–16 · Mintware leads

Seedbase + $CIK
Deploy

Seedbase is Mintware's product — a giving and funding platform where reputation is everything. The agentic layer is Mintware's to build, extending the personal agent's giving capability into full mission matching, donor reputation scoring, and LP access logic. Seedbase is Base-native: it lives in Base's UX and ecosystem, with Firmamint as the settlement layer underneath via the Basemint bridge. $CIK is reestablished with collective backing — not speculative, backed by economic rights in the Firma Collective. The reputation you build giving through Seedbase is the same reputation that follows you everywhere else in the ecosystem.

Step 4.1
Mintware extends personal agent for Seedbase — reputation-based giving engine
Mintware Agent Layer Reputation
Mintware's agentic layer for Seedbase is built on top of the personal agent architecture — it extends the giving/matching module, it does not replace or rebuild the identity layer. The Seedbase agent does: mission alignment scoring (matching your giving history and declared framework to available missions), donor reputation accumulation (consistency, reliability, LP participation quality), and LP matching (connecting proven donors to funding pools for aligned projects). Mintware owns this module entirely. It writes to the shared reputation registry — every giving action in Seedbase becomes part of your Firma identity.
What transfers to NSA and NSSim: Your Seedbase giving history travels. If you've given consistently to climate resilience missions, your agent surfaces that signal when you start building a solar network state in NSSim. If you've supported governance reform projects, your NSA governance participation gets weighted by your track record. The reputation is portable because it's all the same identity.
Seedbase agent module built Mission matching logic live Donor reputation writing to registry
Step 4.2
$CIK reestablished — token deployed with collective economic backing
Mintware Firmamint Base
$CIK is deployed on Firmamint with real utility: access to Seedbase's LP matching, project funding flows, and reputation-gated features. Its value is not speculative — it is backed by the economic rights structure of the Firma Collective and by the activity of the ecosystem it serves. NSA treasury seeds the initial $CIK/USDC pool via the Basemint bridge. Mintware earns economic rights from every Seedbase transaction. $CIK is Base-native — users interact with it in the Base ecosystem; Firmamint settlement is invisible infrastructure behind it.
$CIK deployed on Firmamint $CIK/USDC pool seeded Seedbase public launch First NSA grants deployed through Seedbase
Layer 5
5
Weeks 10–22 · 4 engineers + contracted sim agent shop

NSSim — Build,
Demo, Grant

The Network State Simulator is the product that brings SRI into the ecosystem and gives NSS and Base their specific, funded modules to build. The personal agent's advisory capability gets extended into the Simulator — when you build a network state in NSSim, your agent advises you based on your specific reputation, your giving history, your declared values, your past builds. It is not a generic tool. It is your agent helping you build something that reflects who you are. The SRI demo is the milestone. The three open grant modules are the invitation.

Step 5.1
Contract NSSim advisory agent shop — tight spec, full IP transfer
Contracted Shop Advisory Agent NSSim
The NSSim advisory agent — the layer that reads territory inputs and produces natural language guidance on climate viability, resource estimates, governance recommendations — is scoped, specced by your senior AI engineer, and contracted to a specialist shop. LangChain or LlamaIndex stack. The spec is written first by your GIS engineer (who defines exactly what data comes in) and your senior AI engineer (who defines exactly what outputs are expected). The shop builds the reasoning layer to that spec. Full IP transfer into the Firma Collective trust. The shop earns economic rights for their contribution. This is not a vague "build us an AI" contract. It is a precise deliverable with a defined input/output contract.
The demo principle for SRI: The NSSim advisory agent outputs are primed on publicly available SRI research methodology before the demo. SRI sees their own research language reflected in the Simulator's outputs — not because we copied their work, but because the model has been oriented toward their analytical framework. That's the hook. Then you show them the data slot where their actual validated data would plug in.
NSSim agent spec written Shop contracted IP transfer terms confirmed
Step 5.2
Build NSSim MVP — map interface, sandbox mode, personal agent integration
Full-Stack Eng GIS Eng Blockchain Eng AI Eng
Four engineers build the Simulator. Full-stack builds the interface. GIS engineer builds the map layer and owns the SRI data API spec. Blockchain engineer deploys the on-chain contracts (Governmint, token launchpad, IP Trust submission). The contracted shop builds the advisory agent layer. The personal agent integration is the differentiating feature: NSSim knows who you are. Your agent carries your reputation, your giving history, your past builds, your declared framework into every Simulator session. The advice it gives is calibrated to you, not to a generic user. That is what no other tool can offer.
Map interface live Sandbox mode functional Personal agent integrated On-chain mode contracts deployed SRI data slot visible and specced
Step 5.3
Demo to SRI — show the data slot, deliver the grant module proposal
SRI NSSim Firma
Three-part SRI demo. First: Sandbox mode working — draw a territory, run the analysis, let the advisory agent's language do the work. Second: show the SRI data slot — exactly where their precision mapping and climate research plugs in, what changes when real SRI data is behind it. Third: the SRI Research Grant module — a defined, scoped, paid piece of the Simulator only SRI can build. You are not asking SRI to believe in a vision. You are showing them a running product with a specific empty module and a grant ready to fund them to fill it.
SRI demo delivered Data API spec shared Grant module proposal delivered NSS governance grant published Base tokenomics sandbox grant published
Layer 6
6
Month 6+ · Partners formalize, FIG launches when backing is verified

Partners Formalize.
FIG Launches
Into a Pool That Exists.

By the time you reach this layer, Firmamint is live, Seedbase is running, NSSim has been demoed to SRI, and three grant modules are funded and in progress. You are not pitching a vision. You are showing partners a live ecosystem with real activity. SRI, NSS/Balaji, and Base/Coinbase all sign against something that already exists. FIG launches last — when SRI has validated the asset base, the USDC pool has depth from Base's LP position, and multiple network states are operational on Firmamint. Nothing about FIG's launch is cold.

Step 6.1
SRI signs — data API contract, SRI Ventures co-invest, FIG legitimization begins
SRI NSSim FIG
SRI signs the data API grant. SRI retains all research IP. SRI Ventures takes a co-invest seat in approved NSSim capital rounds. SRI's government relationships begin the FIG legitimization process — identifying jurisdictions where FIG can be recognized as a sovereign currency instrument backed by SRI-validated productive assets. SRI's validation layer is what gives FIG's backing institutional credibility. Without SRI, FIG's assets are claims. With SRI, they are verified productive capacity.
SRI data API contract signed SRI Ventures co-invest seat confirmed FIG legitimization jurisdictions identified
Step 6.2
Base / Coinbase — Basemint named, USDC pool seeded, Ventures path opened
Base / Coinbase Basemint FIG/USDC Pool
Base formalizes: Basemint bridge is named in all documentation as the official USDC bridge partner. Base seeds the initial USDC liquidity pool as anchor LP. Coinbase Ventures receives the IP acquisition path documentation — portfolio companies have a non-traditional exit option through the collective trust. Base builder tokenomics sandbox grant is funded and published on Base's channels. Coinbase's early LP position earns from every transaction on Firmamint from day one — Seedbase activity, NSSim cohort builds, reserve layer settlements. The pool grows with the ecosystem around their position.
Basemint bridge formally named USDC pool seeded Tokenomics sandbox grant published Ventures IP path shared
Step 6.3
Balaji + NSS formalize — governance grant, Cohort One, silent partner round
Balaji / NSS NSSim FIG Pool Rights
NSS governance designer grant funded. Cohort One announced — first NSS students building network states through NSSim. Balaji's silent partner arrangement signed — Firma PIF backs his personal network state project with no strings, no board seat, no timeline. His economic rights in the FIG/USDC pool infrastructure confirmed as named partner. The intellectual foundation of the network state movement aligns with its infrastructure layer. Balaji doesn't endorse Firma. He builds inside it. That is more powerful than endorsement.
NSS governance grant funded Cohort One announced Silent partner round signed Balaji pool rights confirmed
Step 6.4
FIG issued — verified backing, deep pool, Coinbase as primary market maker
FIG Issuance SRI Validated Coinbase MM
FIG launches when — and only when — SRI has validated the asset base, LiquidAcre land rights are on-chain, and the USDC pool has meaningful depth. The sovereign wealth fund issues initial FIG supply calibrated precisely to verified backing. FIG/USDC market pair launches into a pool that already has depth, with a market maker (Coinbase/Base) who has been inside the ecosystem since Layer 3. Price discovery begins. FIG finds its equilibrium — stable by design, backed by real productive assets, not pegged to anything. Every personal agent in the ecosystem can now hold FIG as the internal currency of the union. Citizens on-ramp through the USDC bridge. Trade between network states settles in FIG through the pool.
Asset base SRI-verified FIG initial issuance FIG/USDC pair live Coinbase primary market maker Personal agents holding FIG Network State Union operational

Every layer builds the foundation the next one stands on. The moral framework before the agent. The agent before Seedbase. The chain before the Simulator. The pool before FIG. Nothing is cold. Everything is earned.

Summary

What You Have
After Each Layer

After Layer 0
A moral framework that belongs to everyone
Unlocks
The entry condition for the network exists. Every agent that deploys after this declares against it. The network has a values layer before it has a chain.
After Layer 1
Legal clarity, identity contracted, Mintware signed
Unlocks
NSA can hold treasury. Contracts can be signed. LightCI is building the identity layer. Mintware is speccing Seedbase. You can move fast without legal ambiguity.
After Layer 2
Every person has an agent tied to their values
Unlocks
Soulbound identity live. Reputation registry accumulating. Moral framework enforcement active. Personal agents initialized from declared values. The network has a trust layer.
After Layer 3
A live chain with real infrastructure
Unlocks
Firmamint mainnet running. USDC flowing via Basemint bridge. NSA treasury live. Every partner conversation now starts with "here is the chain, on-chain, right now."
After Layer 4
Live application, reputation accumulating
Unlocks
Seedbase running on Base. $CIK reestablished. Real giving activity generating real reputation on Firmamint. The collective structure is proven, not theoretical.
After Layer 5
Working Simulator, SRI demoed, grants funded
Unlocks
NSSim demo-ready. SRI has seen their data slot. Three grant modules published with defined specs. Partners are handed specific, paid work — not asked to believe.
After Layer 6
Four partners active. FIG live. Union operational.
Unlocks
SRI, Base, NSS, Balaji all building inside the ecosystem. FIG issued against verified backing into a deep pool. Personal agents holding FIG. Network states operational. The whole stack is live.
The Through-Line
Moral framework → identity → agent → chain → application → ecosystem
The sequence
Each layer is the prerequisite for the next. The values layer before the identity layer. The identity layer before the agent. The agent before the chain. The chain before the applications. The applications before the partners. The partners before FIG.