chore: regenerate fish completions for simulate flags

Author: u9g

autocomplete/fish_autocomplete

@@ -226,6 +226,8 @@ complete -c lk -n '__fish_seen_subcommand_from agent a; and __fish_seen_subcomma
 complete -c lk -n '__fish_seen_subcommand_from agent a; and __fish_seen_subcommand_from simulate' -f -l yes -s y -d 'Skip the source-upload confirmation prompt (required for non-interactive runs that generate from source)'
 complete -c lk -n '__fish_seen_subcommand_from agent a; and __fish_seen_subcommand_from simulate' -f -l view -r -d 'Open a pre-existing simulation'
 complete -c lk -n '__fish_seen_subcommand_from agent a; and __fish_seen_subcommand_from simulate' -f -l export -r -d 'Print the run with run `ID` and its exact per-job chat contexts as JSON. Nothing is run or polled: the run must already be finished'
+complete -c lk -n '__fish_seen_subcommand_from agent a; and __fish_seen_subcommand_from simulate' -f -l baseline -r -d 'Compare failures against the finished run with run `ID`: only scenarios that pass there and fail here fail the exit code. Non-interactive (CI) runs only'
+complete -c lk -n '__fish_seen_subcommand_from agent a; and __fish_seen_subcommand_from simulate' -f -l run-id-file -r -d 'Write the simulation run ID to `FILE` as soon as it is available. Non-interactive (CI) runs only'
 complete -c lk -n '__fish_seen_subcommand_from agent a; and __fish_seen_subcommand_from simulate' -f -l agent-name -r -d 'Run against an already-running agent instead of spawning one locally. Pass the registered `NAME`, or "" to target the project\'s default agent (the one that auto-joins every room). Requires --scenarios.'
 complete -c lk -n '__fish_seen_subcommand_from agent a; and __fish_seen_subcommand_from simulate' -f -l help -s h -d 'show help'
 complete -x -c lk -n '__fish_seen_subcommand_from agent a; and __fish_seen_subcommand_from simulate; and not __fish_seen_subcommand_from audio' -a 'audio' -d 'Simulate speech-to-speech interactions using the agent\'s full audio pipeline'