Implement probe envs

[will-maclean]

With a new library, bugs will be aplenty and ease of debugging will be important. Implement the probe envs as outlined here: https://andyljones.com/posts/rl-debugging.html.

• [x] One action, zero observation, one timestep long, +1 reward every timestep
• [x] One action, random +1/-1 observation, one timestep long, obs-dependent +1/-1 reward every time
• [x] One action, zero-then-one observation, two timesteps long, +1 reward at the end
• [x] Two actions, zero observation, one timestep long, action-dependent +1/-1 reward
• [x] Two actions, random +1/-1 observation, one timestep long, action-and-obs dependent +1/-1 reward