The first commandment in sound intelligence work, as any professional will tell you, is skepticism: not only of what the evidence seems to prove but of what sources claim.

That commandment was flagrantly violated in the embarrassing affair of Orlando Jose Tardencillas Espinoza, the 19-year-old adventurer who was to be the State Department's "smoking Sandinista": proof of heavy Nicaraguan involvement in the El Salvador insurgency.

That was in part because U.S. intelligence professionals played no known role in the Tardencillas production and were as astounded by its disagreeable outcome as everyone else.

"They should at least have given him a lie- detector test," said one intelligence analyst, shaking his head at a private briefing. (Like "Mother," Aaron Latham's fictional counterintelligence chief in "Orchids for Mother," intelligence people are convinced that "most people would sooner lie to God than to a polygraph.")

Young Tardencillas switched his previous tale (that he's been trained in Cuba and Ethiopia and sent to El Salvador by the Managua comandantes, 180 degrees between San Salvador and Washington. He presumably would have flunked his polygraph test, cool customer though he is. And it might have occurred to the plodding mind to wonder, anyway, whether this precocious "freedom fighter" had enjoyed so far-flung an education in revolutionary techniques as he claimed.

Whoever his preliminary interrogators were, they failed to glean timely hints that they were dealing with a slippery witness, one that a good country lawyer would have hesitated to put on the stand before a courthouse jury.

But it would be wrong to conclude from the Tardencillas fiasco that the United States lacks reliable information about the extent of Cuban and Soviet scheming in Central America. Only willful blindness could ignore or discount the intelligence publicized last week in a briefing by John Hughes of the Defense Intelligence Agency. The Cubans are working swiftly to turn Nicaragua into a formidable bastion of regional military power.

American intelligence also has good reason to believe that Yasser Arafat has boasted of sending Palestinian pilots and "revolutionaries" to Nicaragua and El Salvador, and that the Vietnamese, another Soviet surrogate, are sending American firearms. But the extent to which this sort of intelligence should be publicly displayed is the subject of a bloody and still unresolved battle within the administration.

Ultimately, the main question is not whether substantial information exists. It does. The question is what it means. The evidence in hand is consistent, for example, with Nicaragua's professed fear of hostile Somoza counter- insurgents, or of some U.S.-financed plan for covert action.

The latter possibility might be more easily dismissed but for the Nixon administration's covert enterprises in Chile. Contrary to some lingering misimpressions, they were more limited in scope, method and aim than is often supposed. But in view of Salvador Allende's uneasy grip, they were probably ill-advised.

Political pressure to push intelligence information beyond its value is a constant. From day one of the Reagan era, the State Department has been under pressure from the hard right to take a more aggressive and militarized line in Central America. In the transition period, this predisposition was signaled by the disgraceful and vindictive hounding of U.S. Ambassador Robert White. That professional diplomat insisted, to the displeasure of Sen. Jesse Helms (R-N.C.) and others, that the situation of El Salvador was politically complex.

There is a correspondingly simplistic view among liberals that intelligence is invariably distorted by ideological or military bias. Often it is. In the Vietnam War period, there were entirely too many "captured documents," too many fudged military estimates, too many dodges and prevarications.

For instance, there is the disturbing story of Anthony Riccio, a CIA analyst who was asked for information about North Vietnamese militia arms at Son Tay, the POW camp ultimately raided by U.S. forces in 1970. When Riccio checked the files and found no information, his boss, irritated, sat down and wrote the answer: "old Japanese and French rifles, pitchforks and machetes." (In fact, the guards had submachine guns.)

How strong the pressure is now to strain available intelligence is unknowable. But the Tardencillas affair is a warning that it may be considerable. Even if the young Nicaraguan had sung the tune he was supposed to sing, one man's war story is not enough to make a case.

Intelligence professionals can produce-- have, indeed, already produced--impressive facts about the hostile penetration of Nicaragua and El Salvador. They have done their duty. For whatever conclusions and policies are based on these facts, and above all for conclusions that outrun them, the administration is strictly accountable.

It is the judgment of politicians, not intelligence professionals, that should be under the closest scrutiny today.