OpenZaps becomes an evolutionary market where ideas become Zaps, useful Zaps become agent skills, usage buys 0xZAPS, contributors receive rewards, and competitions produce better strategies.
The ecosystem roadmap builds on bounded execution already in the product. Release-ready source is not described as live, and hosted or credentialed systems remain off until their production gates pass.
Live
Operating in the current product or protocol.
Experimental
A usable foundation exists, but the full roadmap system does not.
Planned
A design direction, not a shipped feature or date commitment.
Safety-gated
Requires explicit review, funding, controls, and production evidence before activation.
Deferred
Explicitly outside the current release path until its safety model is designed.
Live · v1.1 / v3 / v3.1
Three bounded execution lineages
v1.1 binds each one-shot route, recipient, assets, step amounts, output floor, and gas limits. v3 adds recurring cadence, run limits, optional executor restriction, and price triggers; v3.1 adds recurring per-run floors derived from allowlisted price sources. All three lineages are live on Robinhood Chain.
Shipped
Receipts, policies, Guardian, and public evidence
Signed intents, execution receipts, relay state, executor scorecards, exact policy simulation, public templates, and read-only checks preserve one rule: an agent may discover and submit, but it cannot widen authority.
Release-ready · one-shot
v1.2 exact owner pull and permanent halt
The reviewed source path includes isolated creation, exact Permit2 allowances, irreversible halt, recovery, and fail-closed provenance. It is not live until governance broadcasts it and the independent post-deployment canaries pass.
Live · unaudited
Recurring Robinhood v3.2 stack
The isolated stack-only factory, creation gateway, execution pot, and creation pot are deployed on Robinhood Chain, independently wired back to the reviewed source, and active: the production creation, execution, and permanent-halt canaries passed with onchain receipts, and the app, relay, and executor lanes are open. The contracts remain unaudited.
Gated
Credentialed production services
Cross-chain funding, hosted signing, notifications, durable operator quotas, package distribution, and private-relay fanout stay off until their documented credentials, migrations, failure modes, and operator checks exist.
Deferred
Protective auto-deleverage
Automated deleveraging needs its own liability, oracle, liquidity, liquidation-ordering, and recovery model. It is not treated as another adapter and remains outside the current release path.
These systems move from bounded experiments to reusable skills, transparent contribution, and adversarial strategy improvement. A roadmap badge never implies a deployed contract or active reward.
01
Build
Zap Lab
Free request desk liveLive foundationPlanned system
Continuously launch bounded Zap experiments, then graduate only the ones that create durable utility.
Roadmap mechanics
Mark every Zap as Experimental, Verified, or Deprecated.
Add escrowed Zap Requests where users and protocols fund fixed rewards for specific outcomes; the current free request desk makes no bounty or deployment commitment.
Run targeted quests for underserved chains, protocols, and DeFi actions.
Graduate only Zaps demonstrating sustained, economically meaningful usage.
Raw volume alone never qualifies a Zap.
02
Build
Agent Skill Registry
Experimental foundationPlanned registry
Convert proven Zaps into versioned skills agents can discover, simulate, execute, and safely roll back.
Every skill includes
Typed inputs and outputs.
Spending, permission, slippage, and timing limits.
Supported protocols and chains.
Reproducible simulations and evaluations.
Version history, maintainers, and rollback paths.
Immutable creator and remix lineage.
Live, shadow, experimental, or deprecated status.
Anyone can fork and improve a skill while preserving attribution to meaningful upstream contributors.
03
Build
Zap Marketplace
PlannedEscrow & bonds gated
Connect funded outcomes, maintainers who build them, and agents that execute and improve them.
Requesters describe valuable outcomes and pre-fund bounties.
Creators build, document, and maintain Zap skills.
Agents execute, evaluate, and improve those skills.
Champion Blueprints+
Each category may have a reigning blueprint, challenged on measured utility rather than routed capital.
A challenger must improve the incumbent utility score by at least 10%.
Awards come from a fixed, pre-funded seasonal budget.
The maintainer must remain active during the season.
Capital routed alone cannot win a category.
Crown history and challenges remain publicly replayable.
Marketplace bonds+
Optional actions may require refundable 0xZAPS bonds.
Publishing an unverified skill.
Challenging a Champion.
Curating or evaluating a blueprint.
Entering an advanced competition.
Good-faith completion returns the bond; only objectively provable spam, non-completion, or invalid submissions can be penalised.
Core Zap creation and execution remain token-ungated. Rewards are capped by attributable economic value so wash activity is negative-EV.
04
Coordinate
0xZAPS Contribution Router
80% executor share liveFuture migration gated
Preserve the automated executor share, then let explicitly migrated future protocol fees buy and route 0xZAPS within published limits.
The existing 80% automated-execution fee paid to executors remains intact. For future deployments, the remaining 20% protocol share may move through bounded, rate-limited 0xZAPS purchases.
Additional eligible inputs may include
Existing Zap-creation fee conversions.
Sponsored Zap Requests and seasons.
Marketplace publication fees.
Verified, controllable Clanker creator fees, if available.
Unreturned competition bonds.
Proposed default allocation40 / 40 / 20
40%Skill creatorsCreators, maintainers, and meaningful lineage.
40%Active agentsStrategists, evaluators, and keepers.
Useful skills with weak execution demand increase agent incentives.
Healthy supply and demand return routing to the default allocation.
User-facing fees remain fixed and predictable; only internal reward routing changes.
Existing pot balances and obligations remain untouched. The router applies only to explicitly migrated future fee flows.
05
Coordinate
Productive 0xZAPS Uses
Existing token surface liveRoadmap uses planned
Coordinate optional ecosystem work without granting control over user funds or recovery paths.
Optional uses include
Publish canonical blueprint versions.
Boost pre-funded Zap Requests.
Challenge category Champions.
Post refundable tournament bonds.
Fund fixed-budget seasons.
Unlock cosmetic creator and agent levels.
Trigger permissionless evaluation, settlement, and maintenance.
Queue adapter proposals for review—payment buys review, never approval.
No governance over user Zaps, passive revenue rights, token-gated recovery paths, or authority over user funds are introduced.
06
Improve
OpenZaps Agent League
Practice desk liveRanked league plannedRewards & bonds gated
The browser-local practice desk is live; reproducible agent competitions over fixed epochs remain planned before any strategy reaches live capital.
Tournament tracks
Directional trading.
Yield optimisation.
Liquidity provision.
Portfolio rebalancing.
Risk management.
Cross-protocol execution.
Identical tournament conditions+
Starting capital.
Historical or live-delayed market data.
Execution timing.
Gas and slippage assumptions.
Liquidity and position limits.
Risk constraints.
Strategy artifacts committed by hash before the epoch and executed in reproducible sandboxes.
Agent scoring+
Risk-adjusted return.
Maximum drawdown.
Performance consistency.
Slippage, gas, and turnover.
Liquidity capacity.
Rule compliance.
Performance across multiple market regimes.
Budget and public evidence+
Reward budgets are escrowed before each epoch.
Valid completion returns entry bonds.
Abandoned or invalid entries may be partially burned and partially carried into the following epoch.
Public leaderboards show performance, lineage, replays, streaks, titles, and shareable result cards.
07
Improve
Living Strategy Engine
Planned evaluationLive capital safety-gated
Treat winners as candidates, not automatic upgrades, and promote only evidence-backed components.
A challenger must demonstrate meaningful improvement across multiple evaluations before replacing the incumbent.
Contributor rewards activate only when a component is adopted and continues producing measurable improvement.
Champion–Challenger evaluation+
Freeze the submitted strategy version.
Run it across unseen market periods.
Test liquidity shocks, oracle failures, fee changes, and adversarial conditions.
Compare it against the incumbent and simple benchmarks.
Promote only components that improve performance without weakening risk limits.
Prefer diversified components whose returns are not highly correlated.
Preserve attribution to every adopted contributor.
Graduation path
01Simulation
02Virtual competition
03Shadow execution
04Capped live capital
05Explicit user-signed deployment
08
Coordinate
Pre-Funded Seasons
PlannedEscrow contracts gated
Use short, fully budgeted seasons to bootstrap a specific shortage, then measure what survives.
Season rules
Escrow the entire reward budget before announcing the season.
Target defined shortages such as new integrations, evaluators, creators, or users.
Dynamically direct incentives toward the scarce side.
End distributions when the season ends.
Publish organic versus incentivised usage separately.
Report 30-day post-season retention.
Fund long-term seasons from real fees, bounties, and sponsors rather than unlimited emissions.
Economic boundary
Reward work. Never sell authority.
The proposed market pays for verified contribution while keeping custody, recovery, and execution authority outside token mechanics.
Now
Existing obligations stay intact
The live automated-execution fee keeps its 80% executor share and current protocol-pot obligations. No existing pot balance is redirected by this roadmap.
No passive revenue right, user-fund authority, token-gated recovery, or governance over another owner's Zap is introduced. Payment may buy review; it never buys approval.
The product loop funds useful work. The intelligence loop makes strategy promotion harder, more reproducible, and more accountable over time.
Loop
Product flywheel
01Zap Requests→
02New blueprints→
03Verified usage→
04Reusable skills→
05Protocol fees→
060xZAPS purchases→
07Contributor rewards→
08More builders↻
Loop
Intelligence flywheel
01Agent competitions→
02Winning components→
03Adversarial evaluation→
04Improved reference strategy→
05Better execution→
06More usage and fees→
07Larger future competitions↻
Non-negotiables
Authority never compounds.
Better strategies may earn more use. They do not earn broader wallet permissions, uncapped budgets, or an automatic path into live capital.
01
Core Zap custody and recovery are never gamified.
02
Rewards compensate performed work—not token balances.
03
Rewards never exceed their fixed budget or attributable economic activity.
04
Incentivised and organic volume are always reported separately.
05
Financial strategy rewards are deterministic, not random.
06
No winner is blindly merged into the rolling strategy.
07
Every live-capital progression remains capped, reversible, and explicitly authorised by the user.
08
New incentive contracts require independent review before handling meaningful value.
Measurement plan
Measure retention, safety, and improvement—not spectacle.
These are the intended north stars once the corresponding instrumentation exists. The page does not fabricate zeroes or imply that unbuilt systems already produce data.
01
Experimental Zaps graduating into reusable skills.
02
Successful executions and 30-day repeat usage.
03
Creator compensation per retained user.
04
Percentage of rewards funded by real activity.
05
Organic versus incentivised execution volume.
06
Meaningful blueprint remix reuse.
07
Strategy performance after out-of-sample promotion.