Update guardian output schema (#17061)

## Summary
- Update guardian output schema to separate risk, authorization,
outcome, and rationale.
- Feed guardian rationale into rejection messages.
- Split the guardian policy into template and tenant-config sections.

## Validation
- `cargo test -p codex-core mcp_tool_call`
- `env -u CODEX_SANDBOX_NETWORK_DISABLED INSTA_UPDATE=always cargo test
-p codex-core guardian::`

---------

Co-authored-by: Owen Lin <owen@openai.com>

codex-rs/app-server/src/config_api.rs

@@ -564,7 +564,7 @@ mod tests {
             allowed_web_search_modes: Some(vec![
                 codex_core::config_loader::WebSearchModeRequirement::Cached,
             ]),
-            guardian_developer_instructions: None,
+            guardian_policy_config: None,
             feature_requirements: Some(codex_core::config_loader::FeatureRequirementsToml {
                 entries: std::collections::BTreeMap::from([
                     ("apps".to_string(), false),
@@ -675,7 +675,7 @@ mod tests {
             allowed_approvals_reviewers: None,
             allowed_sandbox_modes: None,
             allowed_web_search_modes: None,
-            guardian_developer_instructions: None,
+            guardian_policy_config: None,
             feature_requirements: None,
             mcp_servers: None,
             apps: None,
@@ -734,7 +734,7 @@ mod tests {
             allowed_approvals_reviewers: None,
             allowed_sandbox_modes: None,
             allowed_web_search_modes: Some(Vec::new()),
-            guardian_developer_instructions: None,
+            guardian_policy_config: None,
             feature_requirements: None,
             mcp_servers: None,
             apps: None,