From 1c597fa2aa9c54ad3097d610fadacfa703c02545 Mon Sep 17 00:00:00 2001 From: Elmehdi Aitbrahim Date: Sat, 15 Aug 2026 18:54:37 -0400 Subject: [PATCH] docs: record the open-source direction in the status review MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The review was written before the direction was decided, so its roadmap listed four engineering issues and ended on an open strategic question. Both are now stale: the question has an answer, and the roadmap has two tracks. WHAT THE MEASURED RESULT ACTUALLY IMPLIES. Every reachable configuration has been measured and none is viable at this venue's cost. That does not mean the project has nothing; it means the strategies were never the asset. The rails are -- nineteen deterministic gates, attestation-based screening that fails closed, and §65.4 qabd encoded as an executable check. There is no other open, auditable Shariah screening engine for spot trading. keel is therefore positioned as that engine, with a reference agent attached to demonstrate it. THE GOVERNANCE QUESTION IS ANSWERED BY THE ARCHITECTURE, and the page now says so where a reader will look for it. Contributors from different madhāhib will disagree about rulings, and ungoverned that fragments a project faster than any technical dispute. keel does not issue rulings: attestation is a human input that fails closed, nothing is derived from price data, and the ruling lives in the attestation rather than the code. A Hanafi and a Shafi'i operator can both use it without the project taking a side. ROADMAP, now two tracks: Track 1, engineering -- the same four issues, but each now shows its milestone. #259 and #260 had no milestone at all; they are the remaining known divergences between what a rule computes, what the backtest models and what the venue does, so Phase 9 was created for them and given the two closed defects of the same class that once voided every performance figure here. Track 2, open source -- fifteen issues across three gated phases. Phase 6 is blocking and unglamorous: the repo is public but legally all-rights-reserved, so nobody may contribute and nobody can find it. The two items that matter most are not code: stating the enforcement/issuance boundary, and deciding the scholarly review path honestly -- including the option of saying plainly that no scholarly review has occurred. An overstated claim in this domain destroys the only thing worth having. The launch is a gate rather than a task, for the same reason. Account figures remain withheld, unchanged and re-verified: this repository is public and nothing in the analysis depends on them. Co-Authored-By: Claude Opus 5 (1M context) --- docs/reports/2026-08-16-status-review.html | 61 +++++++++++++++++----- 1 file changed, 47 insertions(+), 14 deletions(-) diff --git a/docs/reports/2026-08-16-status-review.html b/docs/reports/2026-08-16-status-review.html index cb68ac0c..490095b6 100644 --- a/docs/reports/2026-08-16-status-review.html +++ b/docs/reports/2026-08-16-status-review.html @@ -405,37 +405,70 @@

This week

Six pull requests merged. Two fixed live production defects: a catch-up request that exceeded the venue's 350-candle cap and self-wedged — the gap grew daily, so it could never recover — and a discovery filter that hid admissible assets. Both were verified against real production data, not only tests. Four more closed review findings, including one where a change had removed the only test pinning the admission liquidity floor: the floor could have been cut 5× with a fully green suite.

+
+

Direction

+

The decision: open-source the compliance engine

+

Every reachable strategy configuration has been measured and none is viable at this venue's cost. That result reframes what keel is for — and the answer is the part that works.

+ +
+

The strategies are not the asset. The rails are.

+

Nineteen deterministic rails, attestation-based screening that fails closed, and §65.4 qabd encoded as an executable check that halts entries. There is no other open, auditable Shariah screening engine for spot trading — most halal-investing tooling is a closed app asking you to trust its list.

+

So keel is being positioned as an auditable Shariah compliance engine, with a reference agent attached. The agent demonstrates the engine; it was never the product.

+
+ +

The governance question, and why it is already answered

+

Contributors from different madhāhib will disagree about rulings — what counts as riba exposure, how a staking wrapper or a governance token should be classified. Ungoverned, that fragments a project faster than any technical dispute.

+

The architecture answers it: keel does not issue rulings. Attestation is a human input that fails closed — the operator supplies sector, backing and yield with a source and an attributed name, and keel enforces that deterministically. Nothing is derived from price data. A Hanafi operator and a Shafi'i operator can both use it, because the ruling lives in the attestation, not the code.

+
+

Roadmap

-

What's next, in the order it matters

+

Two tracks, nine milestones, sequenced

+

Track 1 — Engineering, four open issues

- #260Honour a rule's conditional entry price Highest value - The executor discards the entry level a rule computed and routes a market order. This is not cosmetic: it doubled pullback_continuation's trade count and dropped its gross profit factor from 0.92 to 0.77 — the strategy takes trades it was designed to decline. + #260Honour a rule's conditional entry price Phase 9 + The executor discards the entry level a rule computed and routes a market order. Not cosmetic: it doubled pullback_continuation's trade count and dropped its gross profit factor from 0.92 to 0.77 — the strategy takes trades it was designed to decline.
- #259Per-asset slippage - One global 5bp constant is applied from BTC to the thinnest listing, so illiquid assets are credited with execution they cannot achieve — always in the favourable direction, by an unmeasured amount. + #259Per-asset slippage Phase 9 + One global 5bp constant from BTC to the thinnest listing, so illiquid assets are credited with execution they cannot achieve — always favourably, by an unmeasured amount.
- #233Capability-based venue visibility - A read-only API key currently looks identical to a working one until an order fails. + #233Capability-based venue visibility Phase 5 + A read-only API key looks identical to a working one until an order fails.
- #198Robinhood prerequisites - The adapter exists behind the broker port; the live execution path is deliberately not wired until its preconditions are met. + #198Robinhood prerequisites Phase 5 + The adapter exists behind the broker port; the live path is deliberately unwired until its preconditions are met.
+

Phase 9 — Execution & backtest fidelity is new, and it groups these with the two closed defects of the same class that once voided every performance figure in the project. That is the honest home for them: all four are divergences between what a rule computes, what the backtest models, and what the venue actually does.

+ +

Track 2 — Open source, fifteen issues across three gated phases

+
+ + + + + + + + +
Sequenced, because each phase gates the next
PhaseOutcomeIssues
6 — Licence & discoverability BlockingThe repo is public but legally all-rights-reserved: no licence, no description, no topics. Nobody may contribute, and nobody can find it.4
7 — Contributor readinessREADME written for a stranger, the documentation standard stated explicitly, the toolchain floor resolved, and a real on-ramp of shallow issues.7
8 — Scholarly grounding & launchThe fiqh basis written down, a review path decided honestly, an Arabic entry point — then announce, never before.4
+
+ +

Two of these are not code and matter most: stating that keel enforces rulings rather than issuing them, and deciding the scholarly review path honestly — including the option of saying plainly that no scholarly review has occurred. In this domain an overstated claim is not a marketing problem; it destroys the only thing worth having.

+ +

The launch itself is a gate, not a task: nothing is announced until every prior box is ticked. A project that attracts attention before it can absorb it dies of that attention, and there is one first impression with this audience.

-

The open strategic question

-

Every reachable configuration has now been measured and none is viable at this venue's cost. That leaves three honest directions, and they are not equivalent:

+

What remains genuinely open

  • Reduce cost. The measured 12× spread between the cheapest and dearest retail venue is larger than any edge this project has found. Venue economics dominate every parameter we could tune.
    • -
  • Raise per-trade edge. A genuinely new technique — the registry is closed, so this is a reviewed core change, not a parameter.
    • -
  • Accept the result. Keel is a working, compliant, well-tested instrument. Continuing to run DCA and accumulate live evidence is a legitimate outcome.
    • +
  • Raise per-trade edge. A genuinely new technique — the registry is closed, so this is a reviewed core change, not a parameter. The most promising family, turnover reduction, was chased to its primary source and closed: its best-evidenced technique is already swept in our grids and produced nothing.
    • +
  • Let the engine be the contribution. Now the stated direction. The compliance work is valuable whether or not any strategy ever clears its cost.
-

The turnover-reduction literature — the most promising family — was chased to its primary source this week and closed: its best-evidenced technique has already been swept in our own grids and produced nothing.