Add rspec-time-guard to catch hanging specs

[apiology]

Problem:
A spec that hangs — on a network call, an infinite loop, an unresponsive external process — blocks the whole CI run indefinitely instead of failing with a clear timeout, since nothing in the suite currently enforces a time limit.

# before: no limit -- a hang here blocks CI until the job's own timeout kills the whole run
it 'handles the YAML gem aliased to Psych' do
  expect(method_stack).not_to be_empty
end

# after: fails in 400s with a clear timeout instead
it 'handles the YAML gem aliased to Psych', time_limit_seconds: 400 do
  expect(method_stack).not_to be_empty
end

Solution:
Adds the rspec-time-guard gem with a 60s global default (spec_helper.rb), plus longer explicit time_limit_seconds: overrides on the handful of specs that legitimately run longer — pure dev/test tooling, no runtime behavior change.

Each non-default limit reflects a real, CI-observed duration: stdlib/gem resolution and YAML/Thor lookups are legitimately slow, not hung, and the YAML/Psych spec timed out in CI at 240.33s against its first 240s limit — a 0.14% margin.