The governance problem is not purely structural — it is motivational. A marketing curator who approves a technically competent campaign that optimises for vanity metrics, creates a treasury dependency, or attracts the wrong participants has done their domain job while failing the network. Outcome accuracy alone cannot catch this. The Curator Charter is the missing layer.
Every curator affirms the charter on-chain at authorization — it is a condition of their role, not an advisory document. The charter is community-authored and ratified via OpenGov referendum. Its content hash is stored on-chain and referenced in every curator authorization transaction. When the charter is amended through a subsequent vote, active curators must re-affirm within a defined window or their authorization lapses. Two Pebbles does not write the charter. The platform provides the binding mechanism; the community provides the content.
These apply to every curator regardless of specialty:
Charter principles are made tangible through a live, multi-dimension signal board on every proposal. Dimensions are derived from the charter's domain lenses and configured per deployment through governance. Curators update their signals as they work; proposers and voters see the full review state in real time — not just a final approved/rejected outcome.
| State | Meaning |
|---|---|
| ◯ Not yet reviewed | No curator with this domain lens has assessed this dimension yet |
| ◔ Under review | A curator is actively evaluating — written reasoning will follow |
| ▲ Concern | A specific issue has been flagged; written reasoning is attached on-chain |
| △ Conditional | This dimension clears if the proposer addresses one identified issue |
| ✓ Clear | Assessed; no material issues found in this dimension |
Every state change requires written reasoning attached on-chain — a curator cannot set a dimension to Concern without saying why. Coverage gaps are also visible: a proposal where Mission Alignment shows "Not yet reviewed" two days before the decision period closes is a signal any voter or curator can act on. The board is stored in pallet-curator as bounded (dimension, state, curator_id, reasoning_hash) tuples. It is not a voting mechanism — it is an information layer.
A marketing curator raises a Concern on the business model dimension: "Revenue model requires ongoing treasury support after grant period — conflicts with the self-sustainability principle." The proposer sees this immediately, revises their model, and the signal updates to Conditional. When the technical curator begins their review, the prior concern and the proposer's response are already on record. By the time voters see the proposal, the signal board reflects the full shape of the review, not just a single recommendation.
Reputation is built retrospectively — scored when projects reach terminal state, not when proposals are approved. This prevents gaming through volume and aligns every curator's financial incentives with actual project outcomes. All curators earn the same flat VerificationRate for milestone work regardless of tier. Monthly retainers are governance-set.
| Tier | Reputation | Capacity | Vote Weight | Bond |
|---|---|---|---|---|
| Junior Curator | 25–49 | 100 pts | 1× | Bond(Junior) |
| Curator | 50–74 | 125 pts | 1.5× | Bond(Curator) |
| Senior Curator | 75–89 | 150 pts | 2× | Bond(Senior) |
| Expert Curator | 90+ | 200 pts | 3× | Bond(Expert) |
DAO curators — collectives of 3–15 specialists — operate under the same tier structure at the organizational level. ZK credentials allow members to prove collective expertise without revealing individual identities. Reputation is collective: all members share the consequences of every decision, creating strong internal incentives for peer oversight.
Token holders who vote consistently build reputation that could earn staking bonuses and platform rewards — depending on what the connected chain chooses to implement with the participation signal — and eventually qualifies them as curator candidates. The pipeline from delegator to expert curator is organic, not gatekept. Expert Voters with strong track records can apply for curator status with their voting history as credible evidence of qualification.
| Tier | Requirements | Signal Tier | Platform Reward |
|---|---|---|---|
| Engaged Voter | 25%+ of proposals, 3+ months | ParticipationSignal(Engaged) | PlatformReward(Engaged) |
| Active Voter | 50%+, 6+ months, 70%+ outcome alignment | ParticipationSignal(Active) | PlatformReward(Active) |
| Expert Voter | 75%+, 12+ months, 80%+ alignment, discussion contributions | ParticipationSignal(Expert) | PlatformReward(Expert) |
Proposer is the natural entry point to the ecosystem. Submitting an idea to a network running Axiongov means entering a supported, collaborative process — expert curator feedback, structured improvement, and a professional recommendation that gives voters something to evaluate. Not a public trial. Not hostile debate. The pathway: proposer → delegator → active voter → expert voter → curator.
On chains that have adopted the reward model through their own governance, delegators who assign voting power across tracks could also earn stacking bonuses up to MaxGovernanceBonus. All amounts are governance-configurable by the connected chain. Axiongov provides the signal; each platform sets its own reward structure. If a connected chain adopts this model through its own governance, the staking return formula could be: Return = BaseStakingRate × (1 + GovernanceBonus), where GovernanceBonus is capped at MaxGovernanceBonus. Axiongov provides the verified participation data; the chain decides whether and how to act on it.
Curators can prove attributes without revealing identity — a PhD without naming the university, 10+ years of experience without disclosing the employer, Expert Voter status without revealing voting history. Credentials are standardised at the Axiongov layer and portable across all connected deployments via ZK proof format. High-calibre professionals can contribute governance expertise without compromising their primary careers or employment agreements.
Milestone-based payments reduce lump-sum risk, but a proposer who abandons after receiving a first tranche still imposes a real cost: curator time already spent, and a funded milestone that delivered nothing. The escrow mechanism closes this gap using the proposer's own approved budget — no additional collateral required.
Each milestone tranche withholds EscrowBuildRate (default 50%) into a pallet-held escrow until the balance reaches EscrowTarget — one full milestone disbursement. Once the target is reached, subsequent tranches release in full. At the final milestone, the proposer receives their final tranche plus the full accumulated escrow released back to them. A good-faith proposer receives every approved dollar.
On abandonment: curators who performed verified work on the abandoned milestone are paid first from escrow at VerificationRate. The remainder returns to the connected chain's treasury. The proposer's on-chain reputation is penalized, affecting future funding eligibility.
A tier-based bond posted at authorization. Determines which tracks a curator is eligible for. Earns staking returns. Returned after a 90-day waiting period on clean exit from curator status.
MIN(ProposalValue × BondCoverageRatio ÷ N_curators, BondCap) where BondCap = 0.5 × MilestoneValue. Denominated in stablecoins or the proposal's own currency — not exclusively DOT. Locked at assignment; released at terminal state. A curator who cannot cover the top-up cannot accept the assignment.
Proposals above LargeStrategicThreshold engage Expert Curators in an advisory board posture over the full project lifecycle — quarterly steering check-ins, pre-proposal development support, and professional network access for the funded team. The bond exposure for multi-year strategic proposals is so large that failure would effectively end a curator's career. That asymmetry is the mechanism.
If a primary slash does not recover the full verified treasury loss, the gap is distributed across the other curators assigned to that proposal. Each has CollectiveCallWindow to contribute voluntarily; automatic on-chain slash applies to those who don't respond. Staying silent about a co-curator's questionable approval now has a direct financial cost. Peer oversight becomes rational self-interest.
Curator costs are borne by the connected chain's treasury — not by Axiongov. ROI = (NetValue_curated − NetValue_baseline) ÷ TotalCost. Business incubator research shows structured oversight reduces project failure rates from 60–80% to sub-20%. Even a 10% efficiency gain at a treasury of meaningful size exceeds all curator costs by multiples, making the system self-funding once that threshold is crossed.
The April 2026 Drift Protocol breach — $285M drained after North Korean group UNC4736 spent six months posing as a trading firm, building trusted relationships, and social-engineering their way to multisig approvals — is the reference threat. The attack was a human-layer exploit, not a code vulnerability. The curator network addresses this with structural and deliberate defenses that raise the cost of that specific playbook.
Axiongov never emits a signal unlocking a full treasury position in one event. Connected chain treasuries receive incremental verified signals, each tied to curator-confirmed delivery. No single compromised approval can drain a position.
Proposals require approval across independently selected curators who share no social surface. Compromising one is not enough. Simultaneously compromising several without triggering anomaly flags is the real bar.
Above LargeTrancheThreshold, Axiongov enforces a mandatory window between curator quorum and signal emission. A socially-engineered approval cannot trigger an instant signal — the community has time to detect and intervene before the connected chain acts.
Large proposals require demonstrable on-chain history from the proposing entity — minimum account age and prior verified activity. A freshly created front company presenting as an established firm fails automatically.
CuratorCorrelationScore in pallet-signal flags curators approving the same proposers significantly faster than peers, or agreeing beyond statistical baseline — the fingerprint of coordinated social infiltration.
pause_all_releases, callable via low-threshold referendum, halts all pending signals while an incident is investigated. Low cost to trigger; the cost of not having it available is potentially catastrophic.
The "open this folder" attack class works by delivering malware through files the reviewer opens in the course of legitimate work. The defense is separating the document review path from the signing path entirely. Standard proposals: hardened container — read-only mounts, no shared volumes with signing environment, destroyed between sessions. Large proposals: dedicated VM with clean snapshots. Treasurer/Root track: physically separate device that has never touched proposer-supplied content. The signing path and document review path must never share the same execution environment.
Signal computation has been verified end-to-end on-chain. Next.js frontend with typed PAPI access and live block subscriptions is deployed. 159 unit tests passing. The architecture is real, not proposed.
The community votes to proceed at each phase gate — continue, iterate, pause, or abandon. No phase expands until the prior one demonstrates measurable results against pre-defined success criteria.
OpenGov 2.0 is the first production deployment of a three-layer architecture that is chain-agnostic by design. Any on-chain treasury that connects to Axiongov receives the same professional review infrastructure, the same reputation system, and the same signal quality — without replicating the underlying pallets. Curator reputation earned on one deployment is portable to others via standardised ZK proofs at the Axiongov layer. Connected chains act on Axiongov’s signals through their own treasuries; Axiongov never holds or distributes funds.
The deeper thesis: decentralized governance fails not because of bad architecture but because of bad incentives for the humans doing governance work. OpenGov 2.0 pays curators for accurate, charter-aligned judgment. It pays voters for informed participation. It penalises poor decisions through retrospective reputation. It bounds losses through escrow and collective liability. Token holders retain full voting authority throughout — what changes is the quality of information they have and the accountability of those who provide expert guidance.
The question is not whether governance needs improvement — it is whether we have the will to implement systematic solutions, and to build them in a way the rest of the ecosystem can adopt. OpenGov 2.0 provides the blueprint and the first deployment. Axiongov provides the signal layer that makes the model portable. Polkadot has the opportunity to lead this evolution — from volunteer effort to professional infrastructure, while preserving the decentralized foundations that make the network valuable.