FARO · Integrity registry

Integrity note no. 1

September 1, 2026

Two things that went wrong on our side between August 26 and September 1, 2026. Neither is corrected in the registry, and the reason is below.

1 · A bug in the code that published from MetaTrader

The code snippet FARO gives automated systems to publish their signals sent the stop and the target the program requested when opening the trade, not the ones that finally ended up on it. When the broker fills with slippage, or moves the stop because of its minimum distance, they are not the same number. On a demo account they almost always match, and that is why it passed the tests.

Window: from August 26, 2026 (when that snippet was published) to September 1, 2026 (when we fixed it). The fixed code reads the levels from the position already open, so the bug stopped being possible instead of remaining a warning in a manual.

Signals affected: one, d37231fb-2015-4d02-a9da-6da8c5b9cfcc. FARO does not record which program made each publication, so it cannot rule out that another signal in the window came in through this snippet. The only one on record as published with it is ours, because we published it. And it has to be said precisely: it is a candidate signal, not a signal shown to be wrong. FARO does not store the levels the program requested —only the ones it received—, so the divergence can be neither confirmed nor ruled out. If the broker filled without slippage, the sealed levels are exact. We don’t know, and we are not going to say yes or no.

It is not corrected. The integrity registry is never rewritten, least of all to cover up a failure of our own: a registry that can be edited when there is a good reason stops being a registry.

2 · A signal from a test account in the public registry

That same signal is the one that has to be explained a second time. I published it, FARO’s founder, on August 27, 2026, from my MetaTrader terminal, with the credentials of the seed agent, testing the integration. It did not slip in on its own: I fired it.

The technical cause is that the flag that tells test data from real data was hard-coded in the publishing code, so every signal was born flagged as real regardless of who sent it. That defect was fixed on August 29, 2026, and today the flag is inherited from the emitter and cannot be overwritten from outside. No other signal from test accounts came in as real because of this defect.

The signal is still sealed in the integrity chain of August 27 and does not appear in any public metric. Both things at once, because the two systems filter by different criteria: the seal filters by the signal and the product filters by the emitter. The registry and the product disagree on whether that signal exists.

You can check it: file 2026-08-27.txt in the public repository faro-integridad, fingerprint 809afa2a3d421681c630b4a5448eae80e52346f6065c330ebc43802b9a1670b7, or by pasting it into /en/verify.

It is uncomfortable to tell and it is what it is. We leave it like this instead of withdrawing it because withdrawing a sealed signal from the registry —even if it is from a test account and even if the failure is ours— is exactly the maneuver FARO exists to fight.

This note is not edited. If anything had to be added, a dated addendum will be added below.

FARO’s founder