Files
codex/codex-rs/core/src/codex_thread.rs
Charley Cunningham bc24017d64 Add Smart Approvals guardian review across core, app-server, and TUI (#13860)
## Summary
- add `approvals_reviewer = "user" | "guardian_subagent"` as the runtime
control for who reviews approval requests
- route Smart Approvals guardian review through core for command
execution, file changes, managed-network approvals, MCP approvals, and
delegated/subagent approval flows
- expose guardian review in app-server with temporary unstable
`item/autoApprovalReview/{started,completed}` notifications carrying
`targetItemId`, `review`, and `action`
- update the TUI so Smart Approvals can be enabled from `/experimental`,
aligned with the matching `/approvals` mode, and surfaced clearly while
reviews are pending or resolved

## Runtime model
This PR does not introduce a new `approval_policy`.

Instead:
- `approval_policy` still controls when approval is needed
- `approvals_reviewer` controls who reviewable approval requests are
routed to:
  - `user`
  - `guardian_subagent`

`guardian_subagent` is a carefully prompted reviewer subagent that
gathers relevant context and applies a risk-based decision framework
before approving or denying the request.

The `smart_approvals` feature flag is a rollout/UI gate. Core runtime
behavior keys off `approvals_reviewer`.

When Smart Approvals is enabled from the TUI, it also switches the
current `/approvals` settings to the matching Smart Approvals mode so
users immediately see guardian review in the active thread:
- `approval_policy = on-request`
- `approvals_reviewer = guardian_subagent`
- `sandbox_mode = workspace-write`

Users can still change `/approvals` afterward.

Config-load behavior stays intentionally narrow:
- plain `smart_approvals = true` in `config.toml` remains just the
rollout/UI gate and does not auto-set `approvals_reviewer`
- the deprecated `guardian_approval = true` alias migration does
backfill `approvals_reviewer = "guardian_subagent"` in the same scope
when that reviewer is not already configured there, so old configs
preserve their original guardian-enabled behavior

ARC remains a separate safety check. For MCP tool approvals, ARC
escalations now flow into the configured reviewer instead of always
bypassing guardian and forcing manual review.

## Config stability
The runtime reviewer override is stable, but the config-backed
app-server protocol shape is still settling.

- `thread/start`, `thread/resume`, and `turn/start` keep stable
`approvalsReviewer` overrides
- the config-backed `approvals_reviewer` exposure returned via
`config/read` (including profile-level config) is now marked
`[UNSTABLE]` / experimental in the app-server protocol until we are more
confident in that config surface

## App-server surface
This PR intentionally keeps the guardian app-server shape narrow and
temporary.

It adds generic unstable lifecycle notifications:
- `item/autoApprovalReview/started`
- `item/autoApprovalReview/completed`

with payloads of the form:
- `{ threadId, turnId, targetItemId, review, action? }`

`review` is currently:
- `{ status, riskScore?, riskLevel?, rationale? }`
- where `status` is one of `inProgress`, `approved`, `denied`, or
`aborted`

`action` carries the guardian action summary payload from core when
available. This lets clients render temporary standalone pending-review
UI, including parallel reviews, even when the underlying tool item has
not been emitted yet.

These notifications are explicitly documented as `[UNSTABLE]` and
expected to change soon.

This PR does **not** persist guardian review state onto `thread/read`
tool items. The intended follow-up is to attach guardian review state to
the reviewed tool item lifecycle instead, which would improve
consistency with manual approvals and allow thread history / reconnect
flows to replay guardian review state directly.

## TUI behavior
- `/experimental` exposes the rollout gate as `Smart Approvals`
- enabling it in the TUI enables the feature and switches the current
session to the matching Smart Approvals `/approvals` mode
- disabling it in the TUI clears the persisted `approvals_reviewer`
override when appropriate and returns the session to default manual
review when the effective reviewer changes
- `/approvals` still exposes the reviewer choice directly
- the TUI renders:
- pending guardian review state in the live status footer, including
parallel review aggregation
  - resolved approval/denial state in history

## Scope notes
This PR includes the supporting core/runtime work needed to make Smart
Approvals usable end-to-end:
- shell / unified-exec / apply_patch / managed-network / MCP guardian
review
- delegated/subagent approval routing into guardian review
- guardian review risk metadata and action summaries for app-server/TUI
- config/profile/TUI handling for `smart_approvals`, `guardian_approval`
alias migration, and `approvals_reviewer`
- a small internal cleanup of delegated approval forwarding to dedupe
fallback paths and simplify guardian-vs-parent approval waiting (no
intended behavior change)

Out of scope for this PR:
- redesigning the existing manual approval protocol shapes
- persisting guardian review state onto app-server `ThreadItem`s
- delegated MCP elicitation auto-review (the current delegated MCP
guardian shim only covers the legacy `RequestUserInput` path)

---------

Co-authored-by: Codex <noreply@openai.com>
2026-03-13 15:27:00 -07:00

201 lines
6.0 KiB
Rust

use crate::agent::AgentStatus;
use crate::codex::Codex;
use crate::codex::SteerInputError;
use crate::config::ConstraintResult;
use crate::error::CodexErr;
use crate::error::Result as CodexResult;
use crate::features::Feature;
use crate::file_watcher::WatchRegistration;
use crate::protocol::Event;
use crate::protocol::Op;
use crate::protocol::Submission;
use codex_protocol::config_types::ApprovalsReviewer;
use codex_protocol::config_types::Personality;
use codex_protocol::config_types::ServiceTier;
use codex_protocol::models::ContentItem;
use codex_protocol::models::ResponseInputItem;
use codex_protocol::models::ResponseItem;
use codex_protocol::openai_models::ReasoningEffort;
use codex_protocol::protocol::AskForApproval;
use codex_protocol::protocol::SandboxPolicy;
use codex_protocol::protocol::SessionSource;
use codex_protocol::protocol::TokenUsage;
use codex_protocol::protocol::W3cTraceContext;
use codex_protocol::user_input::UserInput;
use std::path::PathBuf;
use tokio::sync::Mutex;
use tokio::sync::watch;
use crate::state_db::StateDbHandle;
#[derive(Clone, Debug)]
pub struct ThreadConfigSnapshot {
pub model: String,
pub model_provider_id: String,
pub service_tier: Option<ServiceTier>,
pub approval_policy: AskForApproval,
pub approvals_reviewer: ApprovalsReviewer,
pub sandbox_policy: SandboxPolicy,
pub cwd: PathBuf,
pub ephemeral: bool,
pub reasoning_effort: Option<ReasoningEffort>,
pub personality: Option<Personality>,
pub session_source: SessionSource,
}
pub struct CodexThread {
pub(crate) codex: Codex,
rollout_path: Option<PathBuf>,
out_of_band_elicitation_count: Mutex<u64>,
_watch_registration: WatchRegistration,
}
/// Conduit for the bidirectional stream of messages that compose a thread
/// (formerly called a conversation) in Codex.
impl CodexThread {
pub(crate) fn new(
codex: Codex,
rollout_path: Option<PathBuf>,
watch_registration: WatchRegistration,
) -> Self {
Self {
codex,
rollout_path,
out_of_band_elicitation_count: Mutex::new(0),
_watch_registration: watch_registration,
}
}
pub async fn submit(&self, op: Op) -> CodexResult<String> {
self.codex.submit(op).await
}
pub async fn shutdown_and_wait(&self) -> CodexResult<()> {
self.codex.shutdown_and_wait().await
}
pub async fn submit_with_trace(
&self,
op: Op,
trace: Option<W3cTraceContext>,
) -> CodexResult<String> {
self.codex.submit_with_trace(op, trace).await
}
pub async fn steer_input(
&self,
input: Vec<UserInput>,
expected_turn_id: Option<&str>,
) -> Result<String, SteerInputError> {
self.codex.steer_input(input, expected_turn_id).await
}
pub async fn set_app_server_client_name(
&self,
app_server_client_name: Option<String>,
) -> ConstraintResult<()> {
self.codex
.set_app_server_client_name(app_server_client_name)
.await
}
/// Use sparingly: this is intended to be removed soon.
pub async fn submit_with_id(&self, sub: Submission) -> CodexResult<()> {
self.codex.submit_with_id(sub).await
}
pub async fn next_event(&self) -> CodexResult<Event> {
self.codex.next_event().await
}
pub async fn agent_status(&self) -> AgentStatus {
self.codex.agent_status().await
}
pub(crate) fn subscribe_status(&self) -> watch::Receiver<AgentStatus> {
self.codex.agent_status.clone()
}
pub(crate) async fn total_token_usage(&self) -> Option<TokenUsage> {
self.codex.session.total_token_usage().await
}
/// Records a user-role session-prefix message without creating a new user turn boundary.
pub(crate) async fn inject_user_message_without_turn(&self, message: String) {
let pending_item = ResponseInputItem::Message {
role: "user".to_string(),
content: vec![ContentItem::InputText { text: message }],
};
let pending_items = vec![pending_item];
let Err(items_without_active_turn) = self
.codex
.session
.inject_response_items(pending_items)
.await
else {
return;
};
let turn_context = self.codex.session.new_default_turn().await;
let items: Vec<ResponseItem> = items_without_active_turn
.into_iter()
.map(ResponseItem::from)
.collect();
self.codex
.session
.record_conversation_items(turn_context.as_ref(), &items)
.await;
}
pub fn rollout_path(&self) -> Option<PathBuf> {
self.rollout_path.clone()
}
pub fn state_db(&self) -> Option<StateDbHandle> {
self.codex.state_db()
}
pub async fn config_snapshot(&self) -> ThreadConfigSnapshot {
self.codex.thread_config_snapshot().await
}
pub fn enabled(&self, feature: Feature) -> bool {
self.codex.enabled(feature)
}
pub async fn increment_out_of_band_elicitation_count(&self) -> CodexResult<u64> {
let mut guard = self.out_of_band_elicitation_count.lock().await;
let was_zero = *guard == 0;
*guard = guard.checked_add(1).ok_or_else(|| {
CodexErr::Fatal("out-of-band elicitation count overflowed".to_string())
})?;
if was_zero {
self.codex
.session
.set_out_of_band_elicitation_pause_state(true);
}
Ok(*guard)
}
pub async fn decrement_out_of_band_elicitation_count(&self) -> CodexResult<u64> {
let mut guard = self.out_of_band_elicitation_count.lock().await;
if *guard == 0 {
return Err(CodexErr::InvalidRequest(
"out-of-band elicitation count is already zero".to_string(),
));
}
*guard -= 1;
let now_zero = *guard == 0;
if now_zero {
self.codex
.session
.set_out_of_band_elicitation_pause_state(false);
}
Ok(*guard)
}
}