A "top secret" report on Soviet failures to comply with SALT I is posing new verification problems for President Carter as he tries to revive SALT II and push it through the Senate.

Whatever the outcome of Carter's effort to get the new Strategic Arms Limitation Treaty back on the tracks, the National Security Council report to the Senate Intelligence Committee -- the most comprehensive study of Soviet SALT I performance ever sent Congress -- shows a pattern of apparent cheating that spells trouble on the verification front.

Beyond that, SALT specialists are asking why this factual log of Soviet misconduct should be classified "top secret" when everything in it has long been known to the Kremlin. Some senators, surprised at the specificity of the NSC study, will insist that both Congress and the public be given routine access to proven cases of sharp Soviet practices in the future.

The most shocking exhibit in the July 20 document reveals elaborate Soviet contortions to avoid "dismantling or destruction" of older-model ICBMs as newer missiles have come into the Soviet arsenal. Called to account by the United States more than three years ago, the Soviets are still trying to wiggle out of what the report calls "precise conformity" with the treaty.

U.S. intelligence discovered in the fall of 1975 that new submarine-launched missiles had passed the trigger point set by the treaty for dismantling two types of older ICBMs -- the SS7 and the SS8. So in March 1976, the Ford administration complained to the Soviets, demanding to know why 51 ICBMs had not been dismantled by March 1 as required under the treaty.

The Soviet member of the U.S.-Soviet commission set up to police compliance admitted that 41 of the 51 launchers had not been destroyed but would be by June 1. The U.S. commissioner called this non-compliance a "serious matter."

On June 1, the United States was formally told that the Soviet Union "was in full compliance." That was the first factual misstatement; during the summer of 1976, "discrepancies at a number of launch sites continued to be observed" by U.S. intelligence, the report states.

At the next compliance session in September 1976, the United States again raised these "discrepancies," noting "the failure to remove dismantled fuel storage tanks and missile-transporter vehicles from some dismantled launch sites."

The Soviet commissioner claimed that "all required actions have been completed" -- but agreed once more to "re-check his information."

By April 1977, U.S. intelligence again reported that "some dismantled propellant storage tanks and missile-handling equipment" had not been removed. Without actually charging the Soviets with cheating, the United States said that "there was still a problem with regard to fully satisfying the . . . agreed procedure."

Move now to October 1977, the next meeting of the compliance commission. The United States again charged that "not all the requirements had been fully satisfied at several sites" and again demanded action. The Soviet commissioner promised to report this complaint to Moscow; and the Carter administration, trying desperately to negotiate SALT II, decided against "further pursuit of this matter."

But one year later, in October 1978, the United States was compelled to complain again about Soviet flimflammery, for the same old reason: dismantling or destruction of the illegal sites had still not occurred "in all cases"; "minor discrepancies" still existed at four sites. Gently chiding, the report stated that this was being raised "not as a compliance matter" but because the United States expects "precise conformity" with the treaty.

The issue is still today unsettled. Last April, the United States once again charged that "dismantling or destruction of ICBM launchers . . . were [sic] not complete at certain sites in the U.S.S.R."

Although Carter administration officials claim that its report proves U.S. verification tools are working, skeptics worry that the evidence of multiple Soviet efforts to cheat makes it virtually certain that some of those efforts must succeed.

A bleaker view, held by some intelligence experts, is that one purpose of the long Soviet battle to avoid "precise conformity" with SALT I has been to test how good U.S. verification procedures really are by forcing the United States to tell what it knows about Soviet cheating.

If so, selective cheating under SALT I may have produced new Soviet methods to avoid detection for cheating under SALT II. Despite its careful, almost powder-puff language, that is the unavoidable message of the NSC report.