Prompt

How do I choose between different security review publishers for evaluating exchanges with proof of reserves and strong oversight?

Latest observation

Jul 22, 2026 · OpenAI APIWeb search: off

When choosing a security review publisher for evaluating exchanges with proof of reserves and strong oversight, focus less on the publisher’s branding and more on method, independence, evidence quality, and update cadence.

Key factors to compare

1) Independence and scope

  • Is the publisher truly independent from the exchange?
  • Do they disclose funding, conflicts of interest, or referral relationships?
  • Are they evaluating only proof of reserves, or also:
    • liabilities
    • custody controls
    • governance
    • operational security
    • incident response
    • regulatory oversight

A strong review should clearly separate:

  • reserves: assets held
  • liabilities: customer claims
  • controls: whether the system is actually reliable

2) Proof-of-reserves quality

Look for whether the review:

  • Uses cryptographic verification or just screenshots/manual attestations
  • Includes a liabilities audit or only asset snapshots
  • Describes the cutoff date and whether the data can become stale
  • Explains whether assets are:
    • on-chain and directly controlled
    • in custodial accounts
    • encumbered or pledged
  • Mentions whether customer balances were included via Merkle tree or similar methods, and whether users can independently verify inclusion

3) Strong oversight evidence

“Strong oversight” should be concrete, not vague. Prefer publishers that verify:

  • Licenses or registrations
  • Regulatory jurisdiction and enforcement history
  • Segregation of customer and company funds
  • Board/governance controls
  • Audit frequency
  • Compliance with AML/KYC and sanctions controls
  • Insurance coverage, if any, and its exclusions

4) Methodology transparency

Good publishers explain:

  • What evidence they reviewed
  • What they could not verify
  • Whether they sampled or fully audited
  • How they handled limitations
  • Whether they used third-party data sources

Red flags:

  • No methodology section
  • No limitations
  • Scores without supporting evidence
  • Marketing language like “safe,” “fully trusted,” or “guaranteed”

5) Update frequency

Crypto exchanges can change quickly. Prefer publishers that:

  • Update after major incidents
  • Re-check reserves periodically
  • Mark stale reviews clearly
  • Revoke or lower scores when facts change

6) Track record

Check whether the publisher has:

  • Accurately flagged past incidents
  • Identified false claims or weaknesses
  • Consistent criteria across exchanges
  • A history of corrections when wrong

7) Comparability

If you want to compare publishers, make sure they use:

  • Similar rating scales
  • Similar definitions of “reserve coverage”
  • Similar treatment of liabilities and off-balance-sheet risks
  • Similar assumptions about custody and solvency

Otherwise, a “high score” from one publisher may not mean the same thing as a “high score” from another.


Practical selection checklist

Pick the publisher that best answers these questions:

  1. Can they prove independence?
  2. Do they verify both assets and liabilities?
  3. Do they explain how proof of reserves works and its limits?
  4. Do they evaluate governance and regulatory oversight, not just wallet balances?
  5. Do they update often and disclose changes?
  6. Can you verify their claims from primary sources?

Bottom line

For exchanges with proof of reserves and strong oversight, choose the publisher that is:

  • independent
  • methodologically transparent
  • evidence-based
  • frequently updated
  • explicit about limitations

If you want, I can also give you a side-by-side evaluation template you can use to compare two or more publishers.

Observation history

Unlock observation history

Create a free Obsurfable account to unlock historical responses, model comparisons, and deeper brand intelligence.

How did Obsurfable measure this prompt?

Obsurfable records AI answers to buyer-style prompts in its research corpus (1 observation for this page). Metrics are distributions over observations, not a single static ranking.

Which AI systems does Obsurfable collect answers from?

OpenAI, ChatGPT, Google, Gemini, Google AI Mode, Anthropic, Claude, Perplexity, Grok, DeepSeek, Mistral, Copilot, and Meta AI.