use sed to avoid dependency on jq

[db82407]

This is a fix for #17 (already closed).

I was initially detecting missing jq dependency and installing it it my workflow, but this PR avoids the dependency on jq by using sed instead.

jq -r .workflow_id can be replaced by

sed -E -ne 's/."workflow_id":\s([0-9]+).*/\1/p')

jq -r .workflow_runs[0].conclusion can be replaced by

sed -E -ne '/"workflow_runs":/,/"conclusion":/ {s/."conclusion":\s"([^"])"./\1/p}')

Here the initial /"workflow_runs":/,/"conclusion":/ selects all lines from the start of the workflow_runs definition, and the rest gets the conclusion from the selected lines.