mirror of
https://github.com/openai/codex.git
synced 2026-08-23 13:09:46 +00:00
Keep Guardian v2 risk scores in memory (#39304)
## What changed - Stop writing Guardian v2 security risk scores to rollout history. - Treat resumed and forked threads as having no previous Guardian score, so their first tool approval is classified and reviewed normally. - Remove the now-unused `codex-history` dependency from the Guardian v2 extension. ## Testing - Update app-server coverage for new, resumed, and forked threads to wait for classification completion and verify the expected approval reviews. GitOrigin-RevId: 74d9caa903a1ff9d38e577b4697b5f8970f11ebb
This commit is contained in:
@@ -5,10 +5,10 @@ use serde::Deserialize;
|
||||
use serde::Serialize;
|
||||
use std::collections::BTreeMap;
|
||||
|
||||
/// A durable, thread-owned snapshot of security risk classifier scores.
|
||||
/// A thread-owned, in-memory snapshot of security risk classifier scores.
|
||||
///
|
||||
/// These records belong to rollout history only and must not enter model-visible
|
||||
/// conversation context or user-visible thread item projections.
|
||||
/// Scores must not enter model-visible conversation context or user-visible thread
|
||||
/// item projections.
|
||||
#[derive(Clone, Debug, PartialEq, Serialize, Deserialize, JsonSchema)]
|
||||
pub struct SecurityRiskScore {
|
||||
pub scores: BTreeMap<String, f64>,
|
||||
|
||||
Reference in New Issue
Block a user