I think a distinction needs to be made between the concept of “a spy” in the singular and the much broader concept of wartime intelligence activities by entire networks and agencies. In a strategic-level game like A&A, I don’t see any plausible way of introducing a house rule that models the activities of “a spy” in the singular because that’s not the level at which single spies operate. The only WWII “spy in the singular” who had a major strategic impact on the entire war was Richard Sorge, Stalin’s top spy in Japan, who alerted Stalin in the fall of 1941 that Japan was planning to attack the US and the UK rather than the USSR; this allowed Stalin to transfer his crack Siberian troops westward just in time to save Moscow from the advancing Germans. Sorge’s impact was spectacular, but it was also extremely exceptional; the vast majority of spies “in the field” simply serve to collect bits of information (often quite modest in importance), which they pass along – often by lengthy routes – to the military or civilian agencies that control their operations. Those agencies then collate and analyze the masses of raw data that arrive from their network of field agents and try to extract useful conclusions from that data. On top of that, these agencies not only take into account this “humint” (human intelligence), they also need to correlate it with “sigit” (signals intelligence), which brings us into the related worlds of radio signal interception and code-breaking. Espionnage in WWII, in other words, was an extremely complex business involving vast numbers of people working in a great variety of tasks, with some of these jobs being extremely sophisticated and specialized, while others were very routine and clerical (though they were essential too). Moreover, intelligence isn’t just a matter of learning the other side’s secrets; it’s also about counter-intelligence, which means either preventing the other side from learning your own secrets or finding a way to feed him false information that actually works to your advantage rather than the enemy’s advantage.
In other words, it would make sense to develop a house rule built around the concept of representing in a collective way these wartime intelligence activities, and of translating these activities into modest bonuses of one sort or another, but I don’t think it would fit A&A’s strategic scale to focus on something as tiny as the activities of one spy, or to allow an espionage house rule to have a strategically major impact on the course of the game.