Exchange Reserve Transparency Whitepaper

Effective Proof-of-Reserves, Stage Framework, and Policy Roadmap

Legal disclaimer. This document is an independent transparency and policy analysis. It is not investment advice, a custody recommendation, or an endorsement of any exchange. Public ratings bind to artifacts and verifier output; they do not guarantee solvency or safety.

Executive Summary

When users deposit crypto on a centralized exchange (CEX), they receive an IOU on the exchange ledger—not on-chain coins under their private keys. History has repeatedly shown that displayed balances can be wrong, incomplete, or unbacked. After the collapse of FTX in November 2022, Proof of Reserves (PoR) declarations spread rapidly across major exchanges. That wave improved narrative transparency but did not end custodial risk.

Four judgments that run through this whitepaper

1. FTX did not end CeFi custody risk. Major freezes, hacks, fraud, and trust crises continued from 2023 through 2026 (AAX, JPEX, WazirX, 2139 Exchange, CoinUp, and others).
2. The post-FTX PoR wave concentrated on Layer 1 (disclosure). Many post-FTX victims used unlicensed or fake “exchanges” (Layer 2). Binance/OKX PoR debates and JPEX-style fraud are different risk layers and must not be conflated.
3. PoR narrative does not prevent crisis freezes. Regional exchanges can halt withdrawals within hours under trust shocks (AAX). Without PoR, rumors alone can trigger run expectations (CoinUp). PoR does not replace liquidity, security operations, or governance resilience.
4. Effective PoR is not a panacea, but gaps can be named. Many failed platforms never offered independently verifiable disclosure. Standards should (a) constrain exchanges capable of disclosure, and (b) help users recognize platforms with no credible custody evidence.

Dual-layer risk model

Policy conclusions

• Stage 0 is not effective PoR. Stage 1 is the minimum policy threshold for Layer 1 exchanges.
• Financial audit and PoR audit are dual-track complements—a clean audit opinion does not imply effective PoR (Chapter 5).
• Every public rating must bind artifacts, hashes, verifier output, and explicit UNVERIFIABLE gaps—not marketing slogans.
• Regulators and institutions must address disclosure quality and platform admission together. Focusing only on headline PoR pages leaves Layer 2 untouched.

Stage at a glance

Current industry snapshot (ardmere public evaluation set, mid-2026): OKX reaches Stage 1 on assessed snapshot 508399035 (Jun 2026); Binance, Bybit, Bitget, Gate.io, and HTX remain Stage 0 with documented gaps. See Chapter 10 and snapshot history.

Reader paths

Part I — Where the Problem Comes From

Chapter 1 — The User View: What You Actually Hold on an Exchange

1.1 After you deposit

On-chain assets move into exchange-controlled wallets. Your account balance is a liability entry on the exchange’s books—an IOU. You do not control those coins with your keys until you withdraw. Legally and technically, you are often an unsecured or specially treated creditor, depending on jurisdiction—not a bailor who “owns coins in a vault.”

1.2 Six common misconceptions

1.3 Runs and liquidity

CEX balances feel like demand deposits. In stress, everyone wants out at once. Exchanges may impose queues, “maintenance,” or full halts—sometimes within hours (FTX, AAX). Annual financial audits are too slow for this risk profile; users need shorter-cycle snapshots with public evidence.

1.4 Bankruptcy

In insolvency, users typically join a creditor queue, not a “return my property” line. Customer-asset segregation rules vary by country. PoR does not change bankruptcy law; it provides ex-ante evidence about whether liabilities were backed at a defined snapshot.

1.5 Self-custody and PoR

PoR improves custodial transparency; it does not eliminate the choice to self-custody. Lost private keys have no bankruptcy court. Policy should understand both paths: self-custody shifts key risk to the user; CEX custody shifts IOU and platform risk to the user.

1.6 Closing thought

Users do not need slogans. They need public evidence before a crisis that what the platform owes them is covered by assets the platform controls—evidence outsiders can fetch and re-verify.

Chapter 2 — Industry History: From “Trust the Platform” to “Show Evidence”

Four phases

Timeline (selected)

2013–14  Merkle liability-proof ideas │ Mt. Gox
2016     Bitfinex hack │ socialized loss
2018–19  Coincheck, QuadrigaCX
2020     Early CEX PoR experiments (e.g. Gate.io)
2022.02  Kraken systematic PoR audits
2022.06  3AC collapse → CeFi contagion
2022.11  FTX │ industry-wide PoR surge
2022.11  AAX withdrawal freeze (immediate aftershock)
2023     JPEX (Hong Kong)
2023+    zk-SNARK/STARK PoR at major venues
2024.07  WazirX hack + governance crisis
2024.09  2139 Exchange-style exit scams
2026.06  CoinUp “exit” rumors & trust crisis (unconfirmed shutdown)

Phase IV closes with a pivot to dual-layer risk (Chapter 3.3): PoR marketing spread among Layer 1 headlines while Layer 2 scams continued under the same word “exchange.”

Chapter 3 — Major Cases: What Users Paid

3.1 Before FTX (selected)

3.2 After FTX: failures and “exits” continued

Conclusion to state plainly: FTX did not end CeFi custody risk. After November 2022, users still faced freezes, hacks, fraud, and trust crises—different mechanisms, same user predicament: no independent verification before the event; creditor treatment after.

Type A — Regional CEX: freeze then effective shutdown

AAX (Atom Asset Exchange), from Nov 2022: Hong Kong venue; cited “system maintenance” to halt withdrawals after FTX; later delisted social channels, offices closed; liquidation and criminal investigations alleging large outflows during the “maintenance” period. ~2M users affected per public reports. PoR wave had barely started; users could not verify “no FTX exposure” claims.

Type B — Unlicensed / fraudulent “exchanges”

JPEX (2023): Unlicensed in Hong Kong; offline promotion and high-yield narratives; SFC warnings; withdrawal friction then collapse. No effective PoR—Layer 2 admission failure, not a Binance Stage 0 debate.

2139 Exchange (2024): CONSOB blocking orders; Ponzi-like yields; mass withdrawal freeze and on-chain outflows. No real trading or reserve program—fraud, not disclosure tuning.

Type C — Security + governance failure

WazirX (from Jul 2024): ~$230M multisig hack; trading/withdrawals suspended; restructuring disputes. Not classic embezzlement, but user outcome resembles a run—brand history did not prevent long freezes.

Type D — Long-tail fake exchanges

Patterns like OrangeX complaints: deposits accepted, withdrawals blocked behind endless KYC/fees—often Layer 2; PoR standards do not audit platforms that never intended to pay.

Type E — Unconfirmed exit / trust crisis (platform still operating)

CoinUp (Jun 2026): Derivatives venue; public warnings involving associated persons; “exit” rumors; CPX token sell pressure; official denials of shutdown and claims of normal deposits/withdrawals. No independently reproducible effective PoR at time of writing; users relied on announcements. Not confirmed as shutdown—classify as trust crisis until regulatory/judicial findings change.

Writing rule: do not label Type E as “confirmed exit” without primary sources; footnote dated press reports.

3.3 Cross-case conclusions: dual-layer risk and PoR limits

The Executive Summary dual-layer table and four judgments apply across Types A–E. Case-specific emphasis:

• Type A (AAX): PoR narrative had barely started; users could not verify “no FTX exposure” before the freeze.
• Type B (JPEX, 2139): Layer 2 admission failure—no artifact path to Stage 0 debate.
• Type C (WazirX): Effective PoR does not prevent hacks or governance breakdowns.
• Type E (CoinUp): Without reproducible artifacts, official denials cannot substitute for evidence.

Do not treat every post-FTX failure as “PoR failure.” Do not use Layer 2 fraud to dismiss Layer 1 disclosure standards.

Chapter 4 — Rise of PoR: From Ideas to Marketing Language

4.1 Origins (2013–2014)

Merkle-sum trees and early proof-of-solvency research combined liability commitments with reserve evidence, enabling privacy-preserving aggregation.

4.2 Early practice (2020–2022.2)

Venues such as Gate.io and Kraken systematized Merkle-based approaches; adoption remained uneven.

4.3 Post-FTX declaration wave (from Nov 2022)

Merkle and zk-SNARK PoR became standard PR. Technical capability improved; policy-grade reproducibility often did not.

4.4 Six gaps between marketing and verification

4.5 What PoR can and cannot do (preview)

4.6 Transition to Part II

Headline PoR quality remains uneven; fake platforms still acquire users under the word “exchange.” Why PoR audit must complement financial audit → Chapter 5. Operational Stage standards → Chapters 6–8.

4.7 Literature note

Academic proof-of-solvency work dates to ~2015; Vitalik’s 2022 “safe CEX” essay and major-venue zk/Merkle specs mark a second wave. IOSCO, Hong Kong SFC, and others address custody and client-asset rules but do not define an Effective PoR / Stage framework like this document. See Appendix H.

Part II — Standards for Effective Disclosure

Chapter 5 — Why CEXs Need PoR Audit: Dual-Track with Traditional Audit

5.1 Two meanings of “audit”

Colloquial “PoR audit” must not be read as financial statement audit with a different name.

5.2 Six reasons users and regulators need PoR audit

① Withdrawals can happen anytime—runs are fast.
Balances feel like demand deposits; crises unfold in days or hours. Annual audits are misaligned with user timing risk.

② Money sits in many places—“company has assets” ≠ “your coins exist.”
Reserves span hot/cold wallets, staking, earn, affiliates. Financial statement line items do not automatically prove customer-level backing.

③ Technology allows public liability commitments.
Merkle trees and zk systems can commit to total user liabilities while letting each user check inclusion—something traditional audit does not typically expose for public rerun.

④ Crises happen between audit periods.
FTX and Celsius showed months-long windows where a prior clean opinion did not protect customer funds.

⑤ Marketing outruns evidence.
“100% backed,” “audited,” “ZK verified” are often misread as personal safety. PoR audit decomposes slogans into downloadable, rerunnable materials—or marks gaps UNVERIFIABLE.

⑥ Audit process is not publicly rerunnable (complement to ⑤).
Financial audit relies on professional independence and workpapers users never see. Effective PoR adds public evidence for custody-specific claims—not an accusation of audit fraud, but a different trust boundary.

5.3 Summary table

CEXs sell custody and redemption, not an annual report alone. Users and regulators need to know what is owed to customers and whether outsiders can check—that is the purpose of PoR audit.

5.4 Clean financial audit still requires effective PoR

Financial audit asks whether company accounts are fairly stated. Effective PoR asks whether customer liabilities are backed and outsiders can verify. A clean opinion on the former does not imply the latter.

FTX (2022.11): Audit ecosystem and reputation coexisted with customer fund commingling—illustrating why custody needs its own reproducible track.

5.5 Five common misuses

5.6 Report taxonomy (do not conflate)

Chapter 6 — Concepts and Terminology

6.0 Terminology chain

Disclosure → PoR audit / review → Effective PoR (Stage 1+) → Public Stage record

Three data planes (keep separate)

1. Reserve side: wallet_address_list, block heights, wallet_ownership_proof
2. Liability side: global_proof, root/vk/config, inclusion proofs
3. Summary side: reserve ratios, scope, cross-binding to (1) and (2)

6.1 Auxiliary lenses: Gen and E Level

Stage is the primary public label. Gen and E Level are explanatory lenses only—they do not upgrade Stage and must not be marketed as parallel ratings.

Common misread: Gen 2 + E2 ≠ Stage 1. Binance publishes Gen 2–class marketing and partial E2 artifacts (summary, wallet_address_list) but remains Stage 0 when wallet_ownership_proof, global_proof, and vk/config are not permissionlessly available (Chapter 10).

6.2 Verifier verdicts

Independent reruns should use explicit outcomes—never treat silence as approval:

Floor flags (e.g. LOGIN_WALL, NO_WALLET_OWNERSHIP_PROOF) cap Stage or confidence when triggered. See Appendix B.

Chapter 7 — PoR Stage Framework

Philosophy borrowed from L2BEAT Stages: Stage measures trust minimization, not a security seal. Domain differs: CEX reserves, not L2 rollups.

7.1 Stage overview

7.2 Pre-Stage and Stage 0

Pre-Stage (any): No public PoR in 12 months; PoR stopped; marketing only; financial reports without PoR artifacts; on-chain treasury display without liability side.

Stage 0 (all): Ongoing disclosure; snapshot boundaries; at least summary or user inclusion; exportable inclusion if offered; open-source verifier or spec.

Typical Stage 0 blockers: no public wallet_address_list, wallet_ownership_proof, or global_proof; opaque trusted setup; login-walled bundles.

7.3 Stage 1 — minimum effective PoR (all required)

S1-5 emphasis: Independent reviewers, media, regulators, and the public must fetch the same artifact set without an account relationship. VIP-only bundles, reviewer-only packages, or split-view releases cap at Stage 0 (LOGIN_WALL floor)—even if other items pass.

Stage 1 equivalent statement:

Aside from implementation bugs and collusion, a third party that long-term fails to detect forged solvency must rely on the exchange replacing artifacts or reviewer collusion—not on “the exchange says it controls these addresses.”

Reference implementation in public set: OKX snapshot 506872725.

7.4 Stage 2 — trust-minimized PoR (directional target)

No complete industry instance at time of writing. Stage 2 builds on all Stage 1 requirements plus the following hard thresholds:

Stage 2 equivalent statement:

Any user can low-barrier verify inclusion; any observer can verify the canonical root for the period; proof constraints reflect real business risk; the exchange cannot unilaterally rewrite anchored history or extend unproven windows.

Third-party archiving ≠ official Stage 2. Observer anchors prove “this third party saw these artifacts,” not exchange-operated canonical anchoring.

7.5 Business consistency and risk flags

7.6 Gap discipline and confidence

• UNVERIFIABLE ≠ PASS
• Ratings bind to a specific snapshot ID
• Floor rules: weakest link caps Stage or lowers confidence
• Single-user inclusion PASS ≠ exchange-wide solvency PASS

Chapter 8 — Minimum Effective PoR Policy (Actionable Clauses)

8.1 Layer 1 (disclosure track)

• Stage 1 is the minimum for licensing, institutional counterparty lists, and public registry effective PoR claims.
• Stage 0 must be labeled as relying primarily on exchange narrative.
• S1-5 permissionless download is a standalone checklist item for supervisors.

8.2 Layer 2 (admission track)

Stage ratings do not apply. Tools: licensing, warnings, domain/payment blocks, criminal enforcement. User education: no public artifact → treat as no custody guarantee (Chapter 11.5).

8.3 Do not conflate

• JPEX-style fraud does not invalidate Stage framework for Layer 1.
• Binance PoR marketing does not protect users from Layer 2 scams.

Part III — How to Verify

Chapter 9 — Evidence Chain and Independent Verification

9.1 End-to-end flow

Exchange publishes artifacts
  → Content-addressed capture (URL, fetch time, content hash)
  → Independent verifier rerun (PASS / FAIL / WARN / UNVERIFIABLE)
  → Artifact bundle root + verification bundle root
  → Optional on-chain anchor (digest only, not bulk data)
  → Public report + gap list bound to snapshot ID

Each public rating must answer: which snapshot, which artifacts, which verifier version, which gaps remain.

9.2 Step 1 — Permissionless artifact fetch (S1-5 test)

From a clean browser session with no exchange account:

1. Locate the exchange’s public PoR landing page or API index for the current snapshot ID.
2. Attempt direct download of all core artifacts: summary, wallet_address_list, wallet_ownership_proof, global_proof, vk/config.
3. Record HTTP status, file hashes, and fetch timestamp. Any registration, login, cookie gate, API key, ticket, or manual approval → LOGIN_WALL floor (Stage 0 cap).

9.3 Step 2 — Verifier rerun

1. Obtain the exchange’s open-source verifier or published algorithm spec (Appendix H.3).
2. Run reserve-side checks: address list vs on-chain balances at declared block height; wallet_ownership_proof batch verification.
3. Run liability-side checks: global_proof + vk/config locally; cross-bind summary ↔︎ reserve aggregate ↔︎ liability root.
4. If user inclusion_proof is tested, confirm it binds to the same public root—single-user PASS ≠ exchange-wide solvency PASS.

9.4 Step 3 — Record and publish

Bind the assessment to:

9.5 Worked example — OKX Stage 1 snapshot 506872725

Illustrative sequence (mid-2026 public evaluation set):

1. Anonymous fetch: Download summary, wallet_address_list, wallet_ownership_proof, global_proof, and vk/config from OKX public PoR endpoints without logging in.
2. Reserve side: Verify address signatures in wallet_ownership_proof; reconcile listed addresses to on-chain balances at the declared snapshot boundary (expect some WARN/UNVERIFIABLE on exotic chains or omnibus labels).
3. Liability side: Run the OKX open-source global ZK verifier (github.com/okx/proof-of-reserves-v2) against the public bundle; confirm summary Merkle root matches global_proof commitment.
4. Stage call: All S1-1–S1-10 hard thresholds met → Stage 1 / effective PoR Yes, with confidence medium (no official canonical on-chain anchor, monthly cadence vs Stage 2 frequency, business-constraint gaps).
5. What this does not prove: No freeze immunity, no hack protection, no coverage of earn/margin outside documented proof boundary.

9.6 Boundaries

1. Third-party archive anchors prove “this observer saw these artifacts”—not official Stage 2.
2. Ratings are evidence records bound to artifacts, hashes, and verifier output.
3. Coverage gaps (e.g. partial on-chain RPC failures) ≠ proof of insolvency.

9.7 Independent registry role (ardmere)

ardmere does not hold user funds, issue licenses, or perform financial statement audits. Empirical exchange assessments for the public evaluation set are in Chapter 10.

Part IV — Industry State and Path Forward

Chapter 10 — Assessed Exchanges (Public Evaluation Set)

10.1 Scope

This chapter covers exchanges in the ardmere public evaluation set—venues with published transparency assessments and machine-readable evaluation records (exchange-assessment@1 schema). It is not an industry-wide ranking: Kraken, Coinbase, and others are out of scope unless separately assessed.

Exclusion from scoreboard: JPEX, 2139, CoinUp (Layer 2 or Type E)—discussed in Chapter 3, not ranked alongside Layer 1.

Forward-looking note: Stages bind to specific snapshot IDs and assessment dates. A Stage label does not predict future snapshots, operational safety, or solvency.

10.2 Summary table (mid-2026)

10.3 OKX — Stage 1 reference (506872725)

• Passes: Public wallet_address_list, wallet_ownership_proof, global_proof, vk/config; S1-5 permissionless download; global ZK binding reproducible with open-source verifier.
• Blocks Stage 2: NO_CANONICAL_ANCHOR, HIGH_FRICTION_INCLUSION_PROOF, BUSINESS_CONSTRAINT_GAP, HIGH_FREQUENCY_GAP.
• Policy read: Current best public reference for minimum effective PoR among assessed majors—not a safety seal.

10.4 Binance (PR01JUN26)

• Partial S1-5 (Jun 2026): summary and wallet_address_list permissionlessly downloadable via public BAPI; wallet_ownership_proof, global_proof, and vk/config are not → fails S1-5, remains Stage 0.
• Flags: NO_WALLET_OWNERSHIP_PROOF, OPAQUE_TRUSTED_SETUP, UNVERIFIABLE on global ZK stack.
• Upgrade P0: Publish ownership proof + permissionless global proof/vk + trusted-setup transcript (Chapter 11.3).

10.5 Bybit (2025061709)

• Available: Summary-style reserve ratio; user Merkle inclusion replayable on supplied test vector (github.com/bybit-exchange/merkle-proof).
• Missing: Public wallet_address_list, wallet_ownership_proof, global_proof.
• Misread to avoid: “I verified my leaf” ≠ exchange-wide solvency.

10.6 Bitget (202605)

• Available: Summary + optional user inclusion under Gen 1 Merkle scheme.
• Defect: 64-bit truncated SHA-256 in public verifier weakens collision resistance.
• Missing: Public wallet list, ownership proof, global proof.

10.7 Gate.io (20260316)

• Available: Public reserve-ratio summary.
• Blocked: ZK bundle behind login; no public wallet_address_list or wallet_ownership_proof → LOGIN_WALL / NO_GLOBAL_PROOF class gaps.

10.8 HTX (20230910)

• Stale sample: 2023 zk sample verifiable in isolation; no current production public bundle in the evaluation pipeline.
• Missing: Contemporary wallet_address_list, ownership proof, and permissionless global proof for ongoing operations.

Official upstream verifier repositories are listed in Appendix H.3.

Chapter 11 — Roadmap and Policy Recommendations

11.0 Three parallel roadmaps

PoR does not eliminate CeFi risk, but it can turn “we can’t tell” into a gap list. Policy and user education should stop Stage 0 slogans on Layer 1 while helping users spot Layer 2 with no artifacts.

11.1 Regulators: dual-track + phased agenda

Disclosure track (Layer 1): Stage 1 floor in license conditions; label Stage 0; require permissionless artifacts (S1-5); bind review type to artifact scope.

Admission track (Layer 2): Licensing, investor alerts, domain blocks, cross-border cooperation—Stage framework does not rate fraud platforms.

Avoid: PoR page without effective PoR definition; financial audit alone for crypto custody; dismissing PoR because of JPEX; PoR without liability side.

11.2 Institutional allocators and brokers

Counterparty tiers:
- Tier A: Stage 1+, reproducible artifacts, bounded gaps
- Tier B: Stage 0 / material UNVERIFIABLE — limits, shorter review cycle
- Tier C: No artifacts or Layer 2 signals — zero tolerance

Due diligence minimum: download current artifacts; verify S1-5; never equate inclusion with exchange-wide solvency; file PoR Stage beside license and financial audit.

11.3 Exchanges (Layer 1 upgrade path)

P0: wallet_ownership_proof, permissionless global_proof + vk, trusted-setup transparency.
P1: proof boundary docs; machine-readable review artifacts.
P2: historical archives; crisis communication discipline (no “maintenance” masking liquidity).

Align each exchange’s open-source verifier repositories (see Appendix H) with production proof versions.

11.4 Media and independent reviewers

Five questions before publishing: Layer 1 or 2? Ever had reproducible artifacts? Financial or PoR audit? Inclusion or global solvency? Does timeline since FTX support “industry safe”?

11.5 Retail users and self-custody participants

On an exchange you hold an IOU, not keys. PoR helps outsiders check before a crisis whether owed balances were backed—it does not guarantee no freeze, hack, or bankruptcy. Goal: avoid fake platforms, limit exposure on real ones, know when to withdraw.

Three checks before depositing

By amount and horizon

Warning signals while deposited

Slow withdrawals, sudden KYC escalation, hack/restructuring news, PoR publications stopping, rumor cycles without public artifacts (CoinUp pattern)—reduce exposure; do not average down on faith alone.

If something goes wrong

Screenshot balances and announcements; file complaints with regulators where applicable; beware recovery scams demanding fees to “unfreeze”; understand bankruptcy may treat you as a creditor, not asset owner.

Do not assume

Advanced users: read transparency reports for UNVERIFIABLE rows; use inclusion tools only with root binding to public summary; see community guides (Appendix H).

11.6 Industry technology roadmap (12–36 months)

11.7 Jurisdiction notes (non-exhaustive)

See Appendix H.6 for IOSCO, SFC, MiCA pointers.

11.8 Priority matrix

Closing: Effective PoR constrains disclosure and aids recognition—not a CeFi safety cure. Regulators, exchanges, institutions, and users must act on four lines in parallel.

Appendices

Appendix A — Stage 1 Practical Checklist

The Stage 1 requirements S1-1 through S1-10 are defined in Chapter 7.3. Use this appendix as a field checklist when assessing a snapshot.

S1-5 quick test: from a clean browser session with no exchange account, attempt direct download of all core artifacts for the current snapshot. Any registration/API gate → Stage 0 cap.

Appendix B — Machine-Readable Rules

Stage thresholds, evidence levels, and risk flags are expressed as structured rules for automation. Independent reviewers should encode the same logic when building assessment pipelines.

B.1 Stage requirement IDs

B.2 Floor flags (weakest-link caps)

B.3 Verdict and schema

• Verdicts: PASS, FAIL, WARN, UNVERIFIABLE (see Chapter 6.2).
• Public evaluation records use schema ardmere/exchange-assessment@1: snapshot metadata, porStage, gen, evidenceLevel, effectivePoR, stageBlockedReasons, riskFlags, and per-check verifier output.

Appendix C — Artifact Vocabulary

Appendix D — Review Taxonomy

See Chapter 5.6 for report types that must not be conflated (technical verification, AUP, limited assurance, reasonable assurance / financial audit).

Appendix E — Major Case Index

Amounts per public regulatory and press sources; update as investigations conclude.

Appendix F — Dual-Layer Risk Diagram

flowchart TB
  subgraph L1["Layer 1 — Disclosure"]
    CEX[Licensed / semi-regulated CEX]
    ST[Stage framework]
    ART[Artifact replay + gap list]
    CEX --> ST --> ART
  end
  subgraph L2["Layer 2 — Admission"]
    FAKE[Unlicensed fake exchanges]
    ENF[Licensing + warnings + enforcement]
    FAKE --> ENF
  end
  USER[User sees word exchange] --> L1
  USER --> L2

Layer 1 (disclosure): PoR Stage separates effective from marketing disclosure; policy tools include Stage 1 floors, artifact replay, and explicit UNVERIFIABLE gaps.

Layer 2 (admission): Policy tool is licensing, warnings, blocks, and enforcement—not Stage ratings.

Post-FTX PoR marketing primarily addressed Layer 1; many post-FTX victim platforms were Layer 2. See Chapter 3.3 and the Executive Summary dual-layer table.

Appendix G — Public Evidence Archive

Each public assessment should bind:

Historical records must remain addressable after exchanges change landing pages—archivers should store content-addressed copies, not rely on mutable URLs alone.

Appendix H — Industry References and Positioning

Positioning: Academic solvency papers, Vitalik’s 2022 essay, Chamber PoR Practitioner’s Guide, exchange tech specs, and IOSCO custody guidance exist—but no widely adopted public policy whitepaper combines Effective PoR Stage floors, dual-track audit, dual-layer risk, and artifact-bound ratings as this document does.

H.1 Academic and cryptographic lineage

H.2 Industry essays and guides

H.3 Exchange technical specifications (single-venue; not industry policy)

H.4 Regulation (custody and client assets; not PoR Stage frameworks)

H.5 L2BEAT Stages (philosophical analogue, different domain)

H.6 Jurisdiction and policy pointers (non-exhaustive)

End of whitepaper draft.