Hello.
Blockchain systems promise transparent rules, verifiable outcomes, and accountable value movement. In practice, a distribution can still be distorted by coordinated wallet structures, shared funding, hidden collection routes, artificial participation, and weak evidence boundaries.
AIE — Airdrop Integrity is a private institutional system for reconstructing wallet relationships and reviewing allocation evidence before an airdrop, points conversion, eligibility decision, or token distribution is finalized.
It is designed to recover structure from fragmented on-chain activity without converting every unusual pattern into an accusation.
Why This Is Relevant to Arbitrum and Orbit
The 2023 ARB airdrop is not the only distribution-integrity decision in the Arbitrum ecosystem.
Arbitrum and Orbit projects continue to operate:
- incentive and points programs;
- eligibility and allowlist processes;
- ecosystem and grant distributions;
- referral campaigns;
- pre-TGE population reviews;
- token allocations;
- post-campaign integrity investigations.
This thread also contains a completed Arbitrum-native controlled reconstruction using public Arbitrum data. The full frozen campaign-level and wallet-level results are published directly below this opening post.
What AIE Does
AIE provides a complete workflow for:
- reconstructing wallet relationships and coordinated structures;
- identifying common funding, collection, and routing patterns;
- separating hidden coordination from shared services and legitimate common behavior;
- preserving the evidence path behind every cluster-level result;
- reviewing individual wallets inside discovered structures;
- distinguishing supported findings from unresolved and insufficient-evidence cases;
- comparing results against approved historical outcomes;
- simulating allocation consequences before execution;
- generating local reports, hashes, checkpoints, and review records.
The analytical result, the supporting evidence, and the final operator or governance decision remain separate layers.
AIE does not convert structural similarity into an automatic finding of ownership, culpability, or abuse.
Private, Buyer-Controlled Operation
Each standard deployment is designed for:
- one organization;
- one trusted operator;
- one licensed Windows workstation;
- unlimited local runs during the valid license period;
- a fixed three-calendar-month license term;
- buyer-controlled datasets, graphs, reports, and evidence records;
- readable static results after license expiry.
The operator works through a local browser interface. Package verification, analysis, graph construction, simulation, manual review, and reporting remain inside the buyer-controlled environment.
The analytical core does not require:
- RPC access;
- external APIs;
- cloud upload;
- telemetry;
- outbound licensing calls.
Upstream data acquisition may be performed online or offline under the buyer’s control. The data-preparation layer converts accepted exports into a deterministic, signed, and hash-addressed package. The Offline Engine accepts only packages that pass identity, schema, checksum, signature, and provenance gates.
buyer-controlled acquisition
→ verified local package
→ offline analytical engine
→ local evidence and reports
External data may enter the workflow. The analytical decision layer does not leave the buyer’s control.
Campaign-Level and Wallet-Level Review
AIE begins at the distribution level.
It identifies coordinated structures across the declared campaign population and shows where deeper examination is justified.
Selected addresses may then be examined separately through AIE — Wallet Risk Analysis for:
- transaction paths;
- exposure;
- counterparties;
- connected activity;
- funding and collection routes;
- mixer or hidden-flow interaction;
- wallet-specific evidence boundaries.
campaign-wide structural detection
→ frozen wallet selection
→ separate wallet-level examination
This creates a controlled path from campaign-level structure to the evidence available for an individual wallet without collapsing the two analytical questions into one score.
Managed Private Analysis
Organizations may request a bounded private review without purchasing or receiving the software.
Available work includes:
- airdrop integrity review;
- wallet risk analysis;
- transaction-path reconstruction;
- mixer and hidden-flow examination;
- wallet incident reconstruction;
- counterparty review;
- independent review of an existing analytical conclusion.
The client and reviewer agree in advance on:
- scope;
- evidence boundary;
- accepted input package;
- confidentiality;
- deliverables;
- timing;
- fee.
The analysis is performed privately, and the client receives a structured result with supporting evidence and explicit limitations.
Requests may be sent by direct message.
Three Public Verification Cases
AIE will accept three technically suitable public verification cases nominated through this forum.
Suitable cases may include:
- a completed historical airdrop;
- a public incentive or points campaign;
- a publication-authorized controlled dataset;
- another frozen wallet population with a defined comparison surface.
For each accepted case:
- the scope and comparison point are fixed before analysis;
- the submitted dataset is frozen;
- the analytical output is frozen before comparison;
- supported, unresolved, contradicted, and untestable claims are separated;
- the result and permitted supporting evidence are published in this thread.
The cases will come from the forum rather than from an internally selected success list.
Limited Institutional Release
The first institutional allocation will contain 11 individually numbered deployments.
Each deployment is intended for one organization, one trusted operator, and one licensed workstation.
Pricing is not published. Deployment scope and institutional terms are discussed privately.
Completed Arbitrum-Native Verification Case
The thread now contains a completed two-stage Arbitrum case based on a frozen, externally supplied issue population.
Case Classification
EXTERNALLY_SEEDED_POST_SNAPSHOT_CONTROLLED_RECONSTRUCTION
The externally supplied allegation was not treated as official Arbitrum ground truth.
No allegation labels, expected clusters, known-Sybil labels, or post-hoc target values were injected into AIE.
Frozen Source Plane
- 1,290 / 1,290 source-manifest entries verified;
- 0 hash mismatches;
- 12,804 normalized direct relationships;
- 1,658 total analyzed addresses;
- 1,656 / 1,656 seed wallets directly linked to the first disclosed central address;
- 1,608 / 1,656 directly linked to the second disclosed central address;
- 1,608 / 1,656 directly linked to both;
- 48 without a direct second-address relationship in the declared plane.
The densest inclusive six-hour window contained:
- 1,653 unique first-reached seed wallets;
- 2,407 direct relationship rows.
The concentrated post-snapshot distribution pattern was supported. Some stronger external numerical claims did not reproduce exactly in the complete declared evidence plane.
Frozen AIE Result
AIE processed the complete frozen plane with:
- 0 imported labels;
- 0 expected clusters;
- unchanged accepted defaults;
- zero network access during analysis.
The native result produced:
- 157 reviewable graph structures;
- 803 seed wallets assigned to native structures;
- 853 seed wallets unresolved or not natively clustered;
- 13,029 candidate edges;
- 2,244 accepted edges;
- 10,785 rejected edges;
- 647 cluster-evidence records;
- 467 group-evidence records.
No wallet reached the configured native high-risk threshold.
AIE did not convert the central-address pattern into an automatic exclusion list.
Separate Wallet-Level Review
A deterministic ten-wallet panel was frozen only after the AIE output had been frozen.
A separate Arbitrum-compatible ingestion candidate was accepted with:
- 154 / 154 tests passed;
- 10 / 10 analytical-core files byte-identical;
- 0 core additions;
- 0 core removals;
- 0 core mismatches;
- 0 outbound network attempts.
Across the exact frozen panel:
- 10 / 10 wallets contained post-snapshot relationship evidence;
- 0 / 10 contained a pre-snapshot relationship in the exact-panel subset;
- 8 / 8 non-central seed wallets were directly linked to both disclosed central addresses;
- 47 unique native transactions were reconstructed;
- 3 token-transfer observations were retained as supplemental evidence;
- 0 wallet-risk scores were emitted;
- 0 automatic-exclusion recommendations were produced.
The absence of a risk score was an evidence-boundary result, not an absence of analysis. Missing receipts, traces, unrelated counterparties, ownership evidence, and complete independent wallet histories were not converted into synthetic confidence.
Combined Conclusion
The frozen result supports:
- a strong post-snapshot operational-coordination or coordinated-execution structure;
- reproducible central-address relationships;
- deterministic separation between structured and unresolved wallets;
- address-level provenance for the wallet-stage evidence;
- a review path that does not depend on accepting a hidden score.
It does not, by itself, establish:
- ownership or common control;
- authorization or compromise;
- operator identity;
- culpability;
- complete wallet histories;
- pre-snapshot eligibility-generating control;
- a basis for automatic exclusion.
The complete Part I and Part II records, chronology, limitations, and evidence hashes are published in the replies below this opening post.
Completed Blind Evidence Case — Hop Protocol
A completed blind analysis was performed on the public Hop topology corpus.
The objective was to determine how much published structure could be recovered from a neutral relationship plane before the official baseline was authorized or opened.
Corpus
- 43,058 official-universe addresses;
- 6,057,473 retained public relationship rows;
- 183 source files;
- Ethereum, Arbitrum, Optimism, Polygon, and Gnosis represented.
Both analytical outputs were hashed, frozen, and receipted before the official baseline was opened.
Full Hop Protocol result
Run A — Comparator-Aligned Neutral Topology
- 2,249 detected structures;
- 12,919 detected addresses;
- 5,819 / 14,195 official Hop exclusions recovered;
- 40.993308% official-exclusion recall;
- 1,514.500 seconds runtime;
- 12 / 610 official groups with 100% member recall;
- 7 exact full member-set recoveries.
Across all 610 official groups:
- 100% member recall: 12
- 90% to below 100%: 5
- 75% to below 90%: 18
- 50% to below 75%: 94
- below 50%: 341
- missed: 140
The result was substantial but partial. It recovered known structure from a reduced topology plane without presenting that reduced plane as the full analytical depth of AIE.
Run B — Raw-Topology Ablation
Run B is classified as:
VALID_RAW_TOPOLOGY_ABLATION_ONLY
It imported the same 43,058 addresses and 6,057,473 directed relationships, but deliberately removed:
- transaction identity;
- real time;
- amount;
- asset;
- method;
- verified contract role;
- trace facts.
Run B recovered:
- 0 / 14,195 official exclusions;
- 1 three-address structure.
It is an ablation result, not a representation of AIE’s complete native capability.
The contrast between Run A and Run B shows where visible graph topology ends and where richer evidence families begin to matter.
Additional Structures
An additional structure was defined as a deduplicated detected structure with zero member overlap with every official Hop group used in the published baseline.
Under that frozen definition:
- independently confirmed: 0
- topology-validated under preregistered criteria: 421
- unresolved: 579
- insufficient evidence: 2
- contradicted: 0
Separately, 1,248 Run A structures overlapped at least one official group and were excluded from the additional-structure count.
Topology-validated structures are not described as confirmed Sybil groups, confirmed abuse, or proof of common control.
Independent Acceptance
Two independent reviews recomputed:
- the accepted 80 / 80 AIE core identity;
- the blind chronology;
- the complete address-level metrics;
- all 610 group-recovery tiers;
- the Run B ablation classification;
- the additional-structure population rule.
They reported zero material metric mismatches.
No AIE rerun, threshold tuning, input change, baseline rewrite, or external-data substitution occurred during acceptance and reconciliation.
Evidence Availability and Reproducibility
The complete frozen evidence package, manifests, receipts, and independent review materials are available for direct examination on request.
No evidence file is attached to this forum post.
The result can be recomputed from:
- frozen input hashes;
- output manifests;
- run receipts;
- the official Hop baseline;
- published evaluation definitions;
- the frozen additional-structure rule.
Decision Boundary
A visible blockchain can still contain an invisible structure.
AIE is designed to recover that structure, preserve the evidence behind it, and return analytical order to decisions that should remain clean, fair, and verifiable.
The analytical system supplies reviewable evidence.
The organization or DAO retains policy, exclusion, appeal, review, and final decision authority.
AIE — Integrity Research
Private Web3 integrity systems for clean distributions, fair participation, and verifiable on-chain decisions.