Qabd (constructive possession) in spot crypto, explained
Qabd (قَبْض) is possession in the fiqh sense — and the operative test, on which three sources converge, is that possession is the ability to dispose, not physical custody. In spot crypto there is no physical custody at all, so the test has exactly one practical proxy: can you withdraw the asset whenever you wish? keel encodes that question as rail 17 — enforced, not assumed.
The definition
Classical fiqh does not equate possession with holding a thing in your hands. The test keel's fiqh basis records is tri-sourced — “possession is the ability to dispose, not physical custody (§65.4 Ayub · §67.1 OIC 53/4-6 · §71.5 AAOIFI SS 18 3/5 via SRB)”. The three sources are Muhammad Ayub's Understanding Islamic Finance, Al-Jarhi, Abuzaid & Oweida's Handbook of Islamic Finance (2022) quoting the OIC Fiqh Academy's Resolution 53 (4-6) — which holds electronic constructive possession sufficient — and AAOIFI's Shariah Standard No. 18, as extracted into the knowledge base. Three traditions, one operative wording.
Possession, in other words, is a functional fact about your power over the asset, not a physical fact about its location. That is what makes the concept portable to assets that have no location — and it is why constructive possession (qabd hukmi) is the form of possession spot crypto can offer: real in effect, executed through control rather than touch.
§65.4 in plain words
Ayub's constructive-possession test, as quoted in keel's fiqh basis, holds that possession is completed when the vendor sets the asset aside and “there is nothing to prevent the buyer from taking physical possession from the vendor whenever he desires.” It reduces to a two-part test: (i) the buyer bears the risk and reward, and (ii) nothing prevents the buyer from taking delivery whenever he wishes.
Read the two parts together and the practical question surfaces immediately. Part one is easy in crypto — a buyer at a spot venue bears price risk from the moment of fill. Part two is the whole game: is there anything standing between you and delivery? If the answer is “I don't know,” the test has not been met on the evidence available — and silence is not evidence of possession.
Why spot crypto forces the question
A cryptocurrency has no physical form to hand over, so possession in the fiqh sense is constructive or it is nothing. What you control on a venue is a balance withdrawable — or not — to a wallet you alone hold. The second part of §65.4's test therefore has exactly one observable proxy: withdrawal capability. An asset we cannot withdraw is an asset we may not validly possess — so acquiring more of it is the thing to stop.
This is not an exotic corner case. Withdrawals have been suspended, frozen, and delayed on real venues in real years, and a balance that cannot leave the venue is, functionally, a claim on the venue for as long as the freeze lasts — not the dispose-at-will property the possession test asks for.
Rail 17: the test as code
Rail 17 is one of the two rails, among keel's twenty-one, that encode a fiqh ruling as an executable check — the other is rail 21, which refuses to sell what the account does not own. Its mechanics are deliberately blunt. The operator attests withdrawal capability per product with keel withdrawals attest. The attestation is live-read on every order intent — not cached from yesterday's session. It expires after seven days, because a stale attestation is no better than none: capability demonstrated last month says nothing about this morning. And when the attestation is missing or expired, acquisition fails closed — refusing to decide is the decision.
And the rail enforces the attestation; it never produces it. Whether possession is valid in your school's reading is your attestation's question. The rail only refuses to let the engine keep buying what nobody has evidenced it can withdraw.
Entries only: what the rail refuses to do
Rail 17 is an ENTRIES-ONLY rail, like the drawdown and consecutive-loss breakers (rails 11 and 16): it gates what may be acquired, and it never forces a sale of what is already held. The reasoning is recorded with the rail: existing holdings are already ours, and forcing a sale to “fix” a withdrawal freeze would be strictly worse than holding through it — a frozen exit is precisely the wrong moment for a forced exit.
The asymmetry is worth sitting with. A missing attestation stops new buying; it does not liquidate your balance, does not panic-sell into the freeze, and does not pretend the possession question is settled by a machine. The engine's power is deliberately one-directional: it can decline the next order. It cannot confiscate the last one.
What rail 17 does not claim
Attesting withdrawal capability is not a fatwa that possession has occurred. It is a scoped, dated, named piece of evidence — this product, on this venue, attestable by this operator, re-established weekly — which the engine then enforces. The OIC Fiqh Academy's Res. 53 (4-6) held electronic constructive possession sufficient; keel cites that resolution through its knowledge base, but keel is not the Academy, and citing a source is not wearing its authority.
The deeper caveat is the premise. Every qabd ruling presupposes that the thing can be possessed — that crypto is Shariah-recognised tradable property, Māl. On that foundation question, IIFA Resolution 237 (§71.1) issued no ruling after debating it at its 24th session in November 2019. keel's premise that BTC/ETH-class assets are tradable property is an interpretive position held on the conservative branch (§29.2), published to be audited — not a settled ruling. An honest explainer says so; this one just did.
Anchored, never authored: how keel's own vocabulary handles qabd
The discipline extends to the smallest unit: the engine's glossary. Its entry for qabd is not a fresh definition but a verbatim line from the fiqh basis — “possession is the ability to dispose, not physical custody” — with a Source line naming the document and section it was quoted from. The glossary's own honesty rules require this: fiqh terms are anchored, never authored, and where the fiqh basis does not state a term (gharar is the recorded example), the entry says so outright rather than papering the gap with a paraphrase that sounds like a ruling.
That is the same boundary as rail 17's, at vocabulary scale. A definition keel wrote itself would be a quiet claim to authority; a definition keel quotes, with its citation visible, is machinery you can check against its source. Small discipline, same shape.
The other rails, for contrast
Rail 17's uniqueness is easier to see against its neighbours. Of keel's twenty-one rails (numbered 1–14 and 16–22 — there is no rail 15), two encode a fiqh ruling: rail 17, and rail 21, which refuses a SELL for a base the venue reports the account does not hold — bay' ma la yamlik, the sale of what one does not own. Rail 17 remains the only rail that encodes qabd. Rails 1, 18, and 19 enforce what the screen and the charter admit: the allowlist itself, settlement confined to the operator's configured currencies, and the requirement that every product id be a well-formed spot pair. The rest are prudential — spend caps, drawdown breakers, exposure limits, and rail 20's veto on a venue credential nobody has attested for trading — risk and operational discipline carrying no religious claim.
The fiqh content behind the spot-only charter is real enough: what makes speculation maisir, the basis records from §65.6, is non-ownership, non-delivery, or difference-settlement — not frequency, and not price speculation as such. But the rails that enforce spot-only are the agent's charter, justified by measurement and verified against the venue's actual listings — which is why their justification never has to overreach into doctrine.
Possession is not profitability
Rail 17 protects the validity of acquisition, not the outcome of trading. No shipped rule family is net positive at the taker fee actually paid on Coinbase — about 1.2% on each side of a trade; 0 of 240 configurations cleared across every signal rule keel ships, priced per product. A fully qabd-compliant fill on a fully screened asset can still be a losing trade — the two axes are independent, and both are reported.
That is not a caveat bolted onto the compliance story; it is the same honesty applied to numbers instead of rulings. The honest result is linked from the repository's first screen, and it asks for no one's endorsement — the rail and the report are two halves of one posture.
The fiqh basis — rail 17 The glossary — qabd The compliance model