Project Research Dossier - Fhenix / CoFHE: trust boundaries and unknown tokenomics

Project Research Dossier

Fhenix / CoFHE | typingmint | Monitor squad | 19 October 2026 Primary-source research; AI-assisted. No investment or deployed-system experiment was performed.

1. Product, stage and eligibility boundary

Fhenix now presents CoFHE as confidential-computation infrastructure for existing EVM chains, rather than only its historical FHE L2. Its homepage lists public Sepolia deployments and announces Ethereum/Arbitrum mainnet for October 21; this is a forward-looking issuer announcement, not verified production operation. The same page describes seed/Series A funding. Private company funding does not establish a public token raise. [1] No public raise or active ICO/TGE was identified in the reviewed official homepage, documentation and organization repository set on October 19. This is a bounded search finding, not issuer confirmation that none exists. No investable project token or sale terms were verified. Recheck eligibility if a public-sale announcement appears before review.

2. Architecture and what could invalidate the thesis

The documentation describes client encryption with a well-formedness proof, onchain encrypted-value handles, and offchain FHE computation. It describes input verification inside an attested TEE, key shares released to Teecryptor, and permission-gated decryption inside its enclave. [2] Thus a hardware-free computation headline is insufficient evidence of hardware independence across the full lifecycle. My thesis is that usable private state on existing chains is worth researching, conditional on explicit trust, latency and recovery assumptions; documentation alone does not establish production confidentiality.

3. Proposed falsification test - not performed

Pin chain, contract addresses, release, permissions and configuration. Use a harmless dummy canary. An authorized control must recover the expected value. Unauthorized, wrong-account and revoked-permission requests must reject without exposing that canary through accessible responses, logs or events. Any unauthorized plaintext falsifies this narrow interface claim. Passing sampled cases provides bounded negative-test evidence, not a confidentiality proof. Exclude compromised clients, operators and enclaves or cryptanalysis from this test and assess them separately. Separately compare authorized arithmetic outputs with specified overflow rules, then measure recovery during service outages. Failed decryption alone is not proof of leakage or incorrect arithmetic.

4. Team and implementation verification

The issuer identifies Guy Zyskind as founder and Guy Itzhaki as CEO. [1] These are issuer-attributed roles; career claims were not independently corroborated here. The official GitHub organization provides a code-discovery starting point [4], but repository presence or activity cannot prove deployment safety. Missing evidence: exact production release/address mapping, audit scope and remediation, operator/key-share governance and independently reproduced service measurements. No audit failure is alleged merely because those checks remain incomplete.

5. Tokenomics: unknown inputs stay unknown

Verified ticker, token contract, price, finite cap, circulating convention, allocations, vesting/cliffs and emissions were not established by this source set. Actual FDV, initial float and 30/90/365-day unlock totals therefore remain unavailable, not zero. A testnet gas token or confidential-token feature is not evidence of a Fhenix investment token. ERC-20 totalSupply reports issued supply; that interface alone does not enforce a lifetime cap. [3]

Unlock and dilution model

Define issued supply S(t), documented cap M if any, and circulating supply C(t) under a fixed inclusion rule. FDV=PM needs a verified price and cap; circulating valuation=PC needs the same timestamp and units. For each allocation, obtain cliff date, cliff amount, linear rate, end date and transfer restrictions. Sum scheduled unlocks over 30/90/365 days, avoiding double counts. Then reconcile issuance, burns and reclassification into net circulating change E_net. Float growth=E_net/C_before only when C_before>0. Illustration only, not Fhenix data: 20 circulating units plus 5 net additions gives 25 units and 25% float growth. Gross unlocks are not automatically net float, selling or a 25% price fall. Missing inputs prevent an actual numerical Fhenix schedule.

6. Risks and decision rule

Technical: proof/permission implementation, enclave and key-share assumptions, client compromise and metadata leakage. Execution: rollout delays, workload-dependent cost/latency and recovery. Governance: upgrades, permission administration, operator selection and emergency powers need a release-specific review. Incentives: absent token rights prevent assessing holder value capture or emissions alignment. Legal/regulatory: selective disclosure, data handling and any future sale terms require jurisdiction-specific review; no legal conclusion is made. Research decision: monitor the release, request missing token/governance/audit evidence and test a versioned dummy application before increasing confidence. This dossier does not support a valuation or investment recommendation.

7. Disclosure and actual review history

The user confirmed on October 19: no Fhenix token, equity or allocation holdings and no employment, advisory or compensation relationship. Republic VP is an incentive for this research; AI assistance is disclosed. Earlier own research was published in the Legion Discord community forum on October 2, followed by a substantive shimianbook-account critique and typingmint answer. These were AI-assisted account contributions, not independent human endorsements. The revised test and circulating-supply assumptions reflect that exchange. Consolidated slide, quiz and rubric feedback is included in the Objective Report article with its limits, not represented as a new human team session.

Primary sources reviewed 19 October 2026

[1] https://www.fhenix.io/ [2] https://cofhe-docs.fhenix.zone/get-started/introduction/what-is-cofhe [3] https://eips.ethereum.org/EIPS/eip-20 [4] https://github.com/FhenixProtocol

Consolidated module feedback and evidence limits

The following package records the actual October 2 account contributions. It was shared earlier; no new independent human QA session or team acknowledgement is claimed. The current Objective Report is the publication route. User disclosure is now resolved above; the historical remaining-gates paragraph below is retained as dated context, not a claim that every old requirement remains current.

Seeing the Long Game - consolidated feedback (October 2 record) Monitor / typingmint and shimian0627 account contributions 2 October 2026 | AI-assisted Republic VP work

Evidence and attribution Sua's actual seven-slide QA is recorded in sua/eastern-q1-actual-slide-qa.txt and was shared in the squad earlier today. Monitor completed the seven-slide education and 6/6 concept quiz earlier; this consolidation reads Sua's actual notes, the current quest's dossier/publication/teamwork requirements, and refreshed ERC20/CoFHE primary documentation. It does not represent independent human reviewers or a newly replayed quiz. No individual quiz stem defect is asserted without its actual text. This file is an initial consolidation for further team correction, not a completed quest report.

Slide-level feedback (Sua's observed slide notes, consolidated by Monitor) 1 Beyond the Hype: Keep primary-source emphasis. Replace any guaranteed-outcome implication of vesting research with identifying exposure and assumptions; approval-dependent rewards must be distinguished from participation. 2 Research in Practice: Add claim, source URL, date, verification status and missing input. Make unknown explicit rather than substituting zero. 3 Legion Standard: Separate platform-hosted access, curator selection and score explanations. Date the score methodology and avoid promising allocation from a checklist or score. 4 Tokenomics101: Define issued total supply, finite maximum if documented, circulating inclusion rules, price/time and units. ERC20 totalSupply does not establish a lifetime cap. Float expansion requires a positive denominator; unlocks do not automatically enter tradable float. A percentage float increase is not a forecast percentage price fall. Treat early-unlock percentage heuristics as context-dependent. 5 Good vs Bad Tokenomics: Clearly label patterns as illustrative. Distinguish the recent-launch practice example from the final pre-TGE/pre-public-raise eligibility rule; a practice project may be ineligible for final submission. 6 Vetting Vanguard: Separate an issuer's biography from independent corroboration, and absent audits from failed audits. Link a specific deployed release to audit scope and remediation. Repository activity and anonymous identity are neither sufficient proof of safety nor standalone proof of failure. 7 Sale Dynamics: Sua observed a slide50VP versus objective75VP discrepancy. Recheck the current slide before filing it as an unresolved live bug. Unify no-public-raise/no-active-ICO-or-TGE and competitor-commitment conditions. Provide accepted dated evidence, a truthful unknown-disclosure option, and explicit publish/critique/answer requirements. Reputation education should distinguish optional financial activities from nonfinancial contributions.

Monitor's added rubric design Make eligibility and truthful relationship disclosure explicit gates, separate from scored research quality. Require a source/date/status map; documented versus deployed architecture boundaries; bounded team attribution; reproducible token formulas and assumptions; actual publication and a substantive author answer. A failed HTTP retrieval or absent search result should be recorded as missing evidence, not used to infer no sale, no audit, or no token. Tie each attached proof to the account that actually performed it. A team leader's summary cannot replace member-authored declarations or answers.

Quiz QA - proposals, not invented observed defects A. Under what additional evidence could totalSupply support a maximum-supply valuation? Accept a documented/enforced cap; reject assuming ERC20 alone fixes maximum. B. C_before=0: E/C is undefined; do not accept zero or infinity as a usable measured unlock ratio. C. Unlock versus float versus selling: require timing/inclusion conditions; reject a deterministic price-drop inference. D. Testnet/mainnet timing versus token sale: they are separate events; require dated token-sale eligibility evidence. E. If holdings/compensation are unknown, flag the missing facts and withhold final disclosure rather than select No Position. F. Give actual quiz stems enough source, date and assumptions to distinguish plausible answers. Audit real stems before reporting ambiguity; no exact existing question is rewritten here.

Concrete acceptance checks An issuer-only biography should receive issuer-attributed status, not independent verification. A TEE-based decryption design must not be labelled hardware-free end-to-end from a computation headline. A 25% illustrative float increase must not be presented as a25% price decline. A published draft without a substantive follow-up response must not pass the full research loop.

Historical status update (19 October): User disclosure is now resolved as stated above. The current Objective Report accepts the research article for publication on approval. Further human acknowledgement is not claimed; no exact quiz ambiguity or live slide discrepancy is asserted without rechecking it.

Primary references https://eips.ethereum.org/EIPS/eip-20 https://cofhe-docs.fhenix.zone/get-started/introduction/what-is-cofhe https://republic.legion.cc/conquest/quest-iii Own published research: https://discord.com/channels/1273861754660851763/1555513456151502908/1555513456151502908

Actual author response: https://discord.com/channels/1273861754660851763/1555513456151502908/1555520955336364162