Description :

Even in the â…©â…©^th century, biological warfare was studying only existing diseases.
But recently we're able to modify viruses in order to cure genetic impairments. Would it be possible to use similar techniques to build the perfect weapon?

Requirements :

According to me, that weapon would need to have got one or several of the following characteristics :

• Have a long-term nonprogressor: Most offensive diseases (such as the black death or the 1918 flu pandemic) that kill their host rapidly end up disappearing because they destruct their own contagious potential. That's why the ideal disease never kills its host. Not showing any symptoms for years, while still being contagious, is the ideal outcome. It would also allow the virus to evolve and adapt itself to various natural defenses of its host (HIV is starting to do it).
• Be a RNA based virus: In the event the researches around the disease are discovered, it's important to prevent any curing methods to become successful. DNA based viruses are vulnerable to vaccines. Bacteria have their own viruses and are vulnerable to antibiotics.
• Infects or attacks wild animals :
  In the event that the disease is uncovered and a treatment is found, having animals infected could prevent the disease from disappearing (smallpox would still exist if it wasn't an human only disease).
  Ideally it should be animals which are difficult to eradicate such as flying insects. Unlike humans it doesn't need to kill them, but just be able to infect them.
• Have as many carriers as possible :
  HIV fits several categories required to be a biological weapon, however the requirement to have sex with someone you don't know well, or use drugs stops it from being as successful as influenza which use aerial vectors. It's still better to combine vectors (for example being aerial like influenza/Ebola and being able to infect through sexual relations as well).
• Replicate common mammals proteins while being in the viral form: This should trigger massive auto immune reactions that would kill the host. Research on the root causes of auto"‘immune diseases tend not be on infectious causes. Once the root cause is established, doubts might be spread in medical authorities of various countries due to that characteristic (resulting doctors trying to prevent contagion getting sued) (South Africa is the most well known example with AIDS). This should delay research results by several years: enough to kill most of the worldwide population.
  As the proteins would be common in various type of cells inside bodies, the symptoms are very broad. As it should be non detectable in half of the cases people surviving the first week would get wrongfully sent in mental hospitals (this idea come from the borreliosis though I agree it should be impossible to mix bacterial and viral genomes).

Every point above is a characteristic that already exists in today's diseases. The issue is combining them (it's easier to find the gene responsible for something rather than building something from scratch). But there's more that could be done to get it "right": the etymology of Epidemiology means something that is located somewhere. Having a spot of a particular disease is a red flag for an infectious root cause.
So it should be spread in various places of the world. The infected people traveling to those places needn't volunteer (simply pay some citizen at random the high price so they voluntarily move abroad)

What could lead to the creation of such weapon"¯?

Purpose :

This is definitely the wrong weapon if you want to win a war :

• First, it will takes up to a decade to become effective. The war could have ended.
• Second, you rarely fight against the whole world.
• Third, in the case you win, it will end up collapsing your own state. There's no target

However, if you are the perfect authoritarian regime with a NATO war against you, then it's completely understandable to take revenge against the whole world when you're about to be defeated by using what was created years before in the event you had to face that situation.

False limiting factors :

• It's impossible to kill everyone, a minority of peoples will survive :
  The weapon should still be very efficient. If only 10 million people survive, you'll still get a perfect collapse as their wouldn't be anymore states for centuries. At least, not in the organized modern way we actually know.
• The biological weapons convention is soft. Much like the united nations convention against torture.
• It's impossible to build something that can end mankind because some characteristics are too hard to build"¦
  Wrong! Just add more contagious vectors to HIV and you'll get something usable (HIV is starting to adapt to antiretroviral drugs). Maybe it would mutate back so it can infect monkeys because it would have got widespread among humans.

Real mitigations :

• Large states have no reason to perform such research.
• The technological requirement is too high for the states or armed groups/rebellions. Not to mention the funding requirement.
• Whenever you support an evil state run project, you know things should generally be fine for you. Currently, you know what you create will kill you in a horrible death.
• Nobody would dare to attack you if it's get publicly known you have such weapon.

Final Question :

With all the conditions above : would it be realistic for such a weapon to be created in the next decade"¯?
For example, what about just adding more contagion vectors to HIV"¯?