Persist security risk scores in rollout history (#38363)

## What changed

- Add a `SecurityRiskScore` rollout item containing a category and numeric score.
- Persist the item in both thread history modes while excluding it from model context, user-visible thread history, search text, forks, and reconstructed conversation history.
- Re-export the score type from the extension API.

## Testing

- Cover serialization, persistence and loading, thread history projection, session reconstruction, append planning, and memory filtering.

GitOrigin-RevId: 1926fe366aeaa75052708a6da589f45a38eefb52
This commit is contained in:
jif
2026-08-13 12:04:29 +00:00
committed by copyberry
parent 9ed0047a61
commit 72fa74fbc9
32 changed files with 165 additions and 4 deletions
@@ -282,6 +282,7 @@ impl Session {
}
RolloutItem::EventMsg(_)
| RolloutItem::SessionMeta(_)
| RolloutItem::SecurityRiskScore(_)
| RolloutItem::InterAgentCommunicationMetadata { .. } => {}
}
@@ -371,6 +372,7 @@ impl Session {
RolloutItem::EventMsg(_)
| RolloutItem::TurnContext(_)
| RolloutItem::WorldState(_)
| RolloutItem::SecurityRiskScore(_)
| RolloutItem::SessionMeta(_) => {}
}
}
@@ -409,6 +411,7 @@ impl Session {
| RolloutItem::InterAgentCommunication(_)
| RolloutItem::InterAgentCommunicationMetadata { .. }
| RolloutItem::TurnContext(_)
| RolloutItem::SecurityRiskScore(_)
| RolloutItem::EventMsg(_) => {
unreachable!("only world-state replay items are collected")
}
@@ -16,6 +16,7 @@ use codex_protocol::protocol::SessionContextWindow;
use codex_protocol::protocol::SessionMeta;
use codex_protocol::protocol::SessionMetaLine;
use codex_protocol::protocol::WorldStateItem;
use codex_protocol::security_risk::SecurityRiskScore;
use pretty_assertions::assert_eq;
use serde_json::json;
use std::path::PathBuf;
@@ -146,6 +147,32 @@ async fn record_initial_history_reconstructs_typed_inter_agent_message() {
);
}
#[tokio::test]
async fn record_initial_history_ignores_security_risk_scores() {
let (session, _turn_context) = make_session_and_context().await;
let user_item = user_message("visible user input");
let security_risk = SecurityRiskScore {
category: "credential_access".to_string(),
score: 0.92,
};
session
.record_initial_history(InitialHistory::Resumed(ResumedHistory {
conversation_id: ThreadId::default(),
history: Arc::new(vec![
RolloutItem::ResponseItem(ResponseItemEnvelope::new(user_item.clone())),
RolloutItem::SecurityRiskScore(security_risk),
]),
rollout_path: Some(PathBuf::from("/tmp/resume.jsonl")),
}))
.await;
assert_eq!(
raw_history_items(&session.state.lock().await.clone_history()),
vec![user_item]
);
}
#[tokio::test]
async fn record_initial_history_restores_world_state_baseline() {
let (session, turn_context) = make_session_and_context().await;
+2
View File
@@ -3256,6 +3256,7 @@ async fn start_new_context_window_assigns_and_persists_item_ids() {
| RolloutItem::InterAgentCommunicationMetadata { .. }
| RolloutItem::TurnContext(_)
| RolloutItem::WorldState(_)
| RolloutItem::SecurityRiskScore(_)
| RolloutItem::EventMsg(_) => None,
});
assert_eq!(
@@ -3322,6 +3323,7 @@ async fn record_initial_history_assigns_and_persists_id_for_forked_response_item
| RolloutItem::Compacted(_)
| RolloutItem::TurnContext(_)
| RolloutItem::WorldState(_)
| RolloutItem::SecurityRiskScore(_)
| RolloutItem::EventMsg(_) => None,
});
let persisted_item = persisted_item.expect("forked response item should be persisted");