I just picked up a copy of Richmond: The Story of a City by Virginius Dabney. The fourth paragraph of page 70 reads as follows.
Chancellor Wythe lived in Ricghmond for seventeen years. In 1806 the town was profoundly shocked when, at age eighty, he was murdered by his grandnephew, George Wythe Sweeney, who put arsenic in his morning coffee. The old man died in his home, after lingering in agony for two weeks. A young mulatto boy who lived with him also died from the poisoned drink. Sweeney's motive is supposed to have been a desire to acquire his inheritance from his granduncle at once, since he was in serious financial trouble. But Wythe lived long enough to realized what had happened and to add a codicil to his will disinheriting Sweeney. The latter escaped punishment because the only convincing witness against him was a Negro woman whom Wythe had freed and who was his devoted servant. Under a Virginia law, ironically drafted by Wythe himself, a black person could not testify against a white person in court.
I have found sources that corroborate the general narrative there. I haven't been able to find anything that confirms that last sentence though. It would be a very interesting piece of trivia if it is in fact true.
If Dabney's right, Wythe either drafted a law that outlawed slave testimony which was previously legal, or one that reaffirmed and/or codified a prior prohibition. In the first case, he's not right. I'm less confident on the second point, but strongly suspect he's wrong there too.
I can't give you direct confirmation in Wythe's murder case, but my present reading is Thomas Morris' Southern Slavery and the Law and he has a chapter on how the rules of evidence interacted with slavery. The rule was that slaves could not testify against whites, period, but they may be allowed to testify against Indians or blacks provided that testimony wasn't in a case where they risked life or dismemberment.
Morris goes through Virginia's laws in some detail. As of 1692, which seems to be the first the colony legislated on the matter, all testimony not under oath was excluded from trials. By the standard of English law at the time, an oath was backed by the swearer's belief that they would face divine sanction if they swore and lied. As a non-Christian had no such fear, they were not considered capable of swearing an oath. They were not, in a word, oathesworth. Since Virginia's slaves were non-Christian at the time (by presumption and probably also in fact for the great majority) that wrote them out of giving testimony at trial. Enslavers of the time worked hard to keep them non-Christian because white supremacy was not yet so secure that they thought they would be let keep Christian slaves.
A 1705 statute on court procedures further affirmed the religious restriction, but now with more of a clearly racial presumption in play:
that popish recusants convict, negroes, mulattoes, and Indian servants, and others, not being christians, shall be deemed and taken to be persons incapable in law, to be witnesses in any cases whatsoever.
Things change a little in 1723. The rule against slave evidence in every case made it hard to prosecute conspiracies to resist slavery and to punish slaves for other crimes. So that needs to change, lest slaves get away with all kinds of things. Courts gain the power to accept any testimony "with pregnant circumstances". That didn't mean you could just take any black person's testimony. Functionally pregnant circumstances meant that one had to have something like a few whites willing to sign off on it and/or something really obvious like a slave accused of stealing a horse being caught riding that horse. But this is still a rule for slaves giving evidence against other slaves, not whites.
Morris says that Virginia's rules remained unchanged on from this point onward so far as slave evidence goes. If he's right on that, then the story doesn't look very good. The principle going back into the seventeenth century is no slaves get to testify against whites. Its codified before 1700 and only changed thereafter to permit slaves to testify against their fellows. Both the original statute and the 1723 rules come before Wythe is even born, so he can't have drafted them. He might have taken part in some kind of formal revision that tidied the code up without substantively changing it which Morris neglected. That would technically fit with what Dabney writes, but it feels like reaching absent further clarification that he meant that rather than the introduction of a new rule.