Source multi-agent instructions from the model catalog (#38619)

## What changed

- Add model-catalog messages for root and subagent roles, explicit delegation, and delegation hints.
- Resolve role instructions in config, catalog, then bundled-default order, while preserving empty values as an explicit way to suppress fallback text.
- Refresh catalog-provided role and mode instructions when the model changes, and give full-history forks the selected child model's subagent role without retaining the parent's role guidance.
- Keep existing config overrides and reasoning-effort behavior, including proactive delegation for ultra reasoning effort.

## Testing

- Cover message deserialization and preservation through model overrides.
- Cover precedence, empty overrides, model switches, resumed sessions, and full-history subagent forks.

GitOrigin-RevId: 4625cf7c6a5490176adddfaa0fb99100707daea9
This commit is contained in:
rhan-oai authored and copyberry committed 2026-08-14 18:26:20 +00:00
1 parent 14a6813fa8
commit 395723b238
30 files changed
+734 -231

No files matched your search

@@ -1808,6 +1808,7 @@ mod tests {
policy_template: Some(catalog_template.to_string()),
}),
permissions: None,
multi_agent: None,
token_budget: None,
};
@@ -1842,6 +1843,7 @@ mod tests {
policy_template: None,
}),
permissions: None,
multi_agent: None,
token_budget: None,
};
@@ -1884,6 +1886,7 @@ mod tests {
policy_template: Some(String::new()),
}),
permissions: None,
multi_agent: None,
token_budget: None,
};