Refactor: load execution profiles from YAML and split prompt definitions into promptdef repository

internal/profile/testdata/valid_with_api_key_env.yaml

@@ -0,0 +1,7 @@
+id: local-secure
+endpoint: http://localhost:8000/v1
+model: gpt-4o-mini
+api_key_env: SCRIPTORIUM_API_KEY
+reasoning_effort: medium
+extra_params:
+  provider: local