Add goal persistence foundation (1 / 5) (#18073)

Adds the persisted goal foundation for the rest of the stack. This PR is
intentionally limited to feature flag and state-layer behavior;
app-server APIs, model tools, runtime continuation, and TUI UX are
layered in later PRs.

## Why

Goal mode needs durable thread-level state before clients or model tools
can safely build on it. The state layer needs to know whether a goal
exists, what objective it tracks, whether it is active, paused,
budget-limited, or complete, and how much time/token usage has already
been accounted.

## What changed

- Added the `goals` feature flag and generated config schema entry.
- Added the `thread_goals` state table and Rust model for persisted
thread goals.
- Added state runtime APIs for creating, replacing, updating, deleting,
and accounting goal usage.
- Added `goal_id`-based stale update protection so an old goal update
cannot overwrite a replacement.
- Kept this PR scoped to persistence and state runtime behavior, with no
app-server, model-facing, continuation, or TUI behavior yet.

## Verification

- Added state runtime coverage for goal creation, replacement, stale
update protection, status transitions, token-budget behavior, and usage
accounting.
This commit is contained in:
Eric Traut
2026-04-24 20:51:38 -07:00
committed by GitHub
Unverified
parent 8a559e7938
commit 0ee737cea6
8 changed files with 1401 additions and 0 deletions
+6
View File
@@ -424,6 +424,9 @@
"general_analytics": {
"type": "boolean"
},
"goals": {
"type": "boolean"
},
"guardian_approval": {
"type": "boolean"
},
@@ -2616,6 +2619,9 @@
"general_analytics": {
"type": "boolean"
},
"goals": {
"type": "boolean"
},
"guardian_approval": {
"type": "boolean"
},
+8
View File
@@ -185,6 +185,8 @@ pub enum Feature {
DefaultModeRequestUserInput,
/// Enable automatic review for approval prompts.
GuardianApproval,
/// Enable persisted thread goals and automatic goal continuation.
Goals,
/// Enable collaboration modes (Plan, Default).
/// Kept for config backward compatibility; behavior is always collaboration-modes-enabled.
CollaborationModes,
@@ -928,6 +930,12 @@ pub const FEATURES: &[FeatureSpec] = &[
stage: Stage::Stable,
default_enabled: true,
},
FeatureSpec {
id: Feature::Goals,
key: "goals",
stage: Stage::UnderDevelopment,
default_enabled: false,
},
FeatureSpec {
id: Feature::CollaborationModes,
key: "collaboration_modes",
@@ -0,0 +1,11 @@
CREATE TABLE thread_goals (
thread_id TEXT PRIMARY KEY NOT NULL REFERENCES threads(id) ON DELETE CASCADE,
goal_id TEXT NOT NULL,
objective TEXT NOT NULL,
status TEXT NOT NULL CHECK(status IN ('active', 'paused', 'budget_limited', 'complete')),
token_budget INTEGER,
tokens_used INTEGER NOT NULL DEFAULT 0,
time_used_seconds INTEGER NOT NULL DEFAULT 0,
created_at_ms INTEGER NOT NULL,
updated_at_ms INTEGER NOT NULL
);
+5
View File
@@ -44,12 +44,17 @@ pub use model::Stage1JobClaimOutcome;
pub use model::Stage1Output;
pub use model::Stage1OutputRef;
pub use model::Stage1StartupClaimParams;
pub use model::ThreadGoal;
pub use model::ThreadGoalStatus;
pub use model::ThreadMetadata;
pub use model::ThreadMetadataBuilder;
pub use model::ThreadsPage;
pub use runtime::DeviceKeyBindingRecord;
pub use runtime::RemoteControlEnrollmentRecord;
pub use runtime::ThreadFilterOptions;
pub use runtime::ThreadGoalAccountingMode;
pub use runtime::ThreadGoalAccountingOutcome;
pub use runtime::ThreadGoalUpdate;
pub use runtime::logs_db_filename;
pub use runtime::logs_db_path;
pub use runtime::state_db_filename;
+4
View File
@@ -3,6 +3,7 @@ mod backfill_state;
mod graph;
mod log;
mod memories;
mod thread_goal;
mod thread_metadata;
pub use agent_job::AgentJob;
@@ -25,6 +26,8 @@ pub use memories::Stage1JobClaimOutcome;
pub use memories::Stage1Output;
pub use memories::Stage1OutputRef;
pub use memories::Stage1StartupClaimParams;
pub use thread_goal::ThreadGoal;
pub use thread_goal::ThreadGoalStatus;
pub use thread_metadata::Anchor;
pub use thread_metadata::BackfillStats;
pub use thread_metadata::ExtractionOutcome;
@@ -38,6 +41,7 @@ pub(crate) use agent_job::AgentJobItemRow;
pub(crate) use agent_job::AgentJobRow;
pub(crate) use memories::Stage1OutputRow;
pub(crate) use memories::stage1_output_ref_from_parts;
pub(crate) use thread_goal::ThreadGoalRow;
pub(crate) use thread_metadata::ThreadRow;
pub(crate) use thread_metadata::anchor_from_item;
pub(crate) use thread_metadata::datetime_to_epoch_millis;
+109
View File
@@ -0,0 +1,109 @@
use anyhow::Result;
use anyhow::anyhow;
use chrono::DateTime;
use chrono::Utc;
use codex_protocol::ThreadId;
use sqlx::Row;
use sqlx::sqlite::SqliteRow;
use super::epoch_millis_to_datetime;
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum ThreadGoalStatus {
Active,
Paused,
BudgetLimited,
Complete,
}
impl ThreadGoalStatus {
pub fn as_str(self) -> &'static str {
match self {
Self::Active => "active",
Self::Paused => "paused",
Self::BudgetLimited => "budget_limited",
Self::Complete => "complete",
}
}
pub fn is_active(self) -> bool {
self == Self::Active
}
pub fn is_terminal(self) -> bool {
matches!(self, Self::BudgetLimited | Self::Complete)
}
}
impl TryFrom<&str> for ThreadGoalStatus {
type Error = anyhow::Error;
fn try_from(value: &str) -> Result<Self> {
match value {
"active" => Ok(Self::Active),
"paused" => Ok(Self::Paused),
"budget_limited" => Ok(Self::BudgetLimited),
"complete" => Ok(Self::Complete),
other => Err(anyhow!("unknown thread goal status `{other}`")),
}
}
}
#[derive(Debug, Clone, PartialEq, Eq)]
pub struct ThreadGoal {
pub thread_id: ThreadId,
pub goal_id: String,
pub objective: String,
pub status: ThreadGoalStatus,
pub token_budget: Option<i64>,
pub tokens_used: i64,
pub time_used_seconds: i64,
pub created_at: DateTime<Utc>,
pub updated_at: DateTime<Utc>,
}
pub(crate) struct ThreadGoalRow {
pub thread_id: String,
pub goal_id: String,
pub objective: String,
pub status: String,
pub token_budget: Option<i64>,
pub tokens_used: i64,
pub time_used_seconds: i64,
pub created_at_ms: i64,
pub updated_at_ms: i64,
}
impl ThreadGoalRow {
pub(crate) fn try_from_row(row: &SqliteRow) -> Result<Self> {
Ok(Self {
thread_id: row.try_get("thread_id")?,
goal_id: row.try_get("goal_id")?,
objective: row.try_get("objective")?,
status: row.try_get("status")?,
token_budget: row.try_get("token_budget")?,
tokens_used: row.try_get("tokens_used")?,
time_used_seconds: row.try_get("time_used_seconds")?,
created_at_ms: row.try_get("created_at_ms")?,
updated_at_ms: row.try_get("updated_at_ms")?,
})
}
}
impl TryFrom<ThreadGoalRow> for ThreadGoal {
type Error = anyhow::Error;
fn try_from(row: ThreadGoalRow) -> Result<Self> {
Ok(Self {
thread_id: ThreadId::try_from(row.thread_id)?,
goal_id: row.goal_id,
objective: row.objective,
status: ThreadGoalStatus::try_from(row.status.as_str())?,
token_budget: row.token_budget,
tokens_used: row.tokens_used,
time_used_seconds: row.time_used_seconds,
created_at: epoch_millis_to_datetime(row.created_at_ms)?,
updated_at: epoch_millis_to_datetime(row.updated_at_ms)?,
})
}
}
+5
View File
@@ -20,6 +20,7 @@ use crate::apply_rollout_item;
use crate::migrations::runtime_logs_migrator;
use crate::migrations::runtime_state_migrator;
use crate::model::AgentJobRow;
use crate::model::ThreadGoalRow;
use crate::model::ThreadRow;
use crate::model::anchor_from_item;
use crate::model::datetime_to_epoch_millis;
@@ -58,6 +59,7 @@ mod backfill;
mod device_key;
#[cfg(test)]
mod device_key_tests;
mod goals;
mod logs;
mod memories;
mod remote_control;
@@ -66,6 +68,9 @@ mod test_support;
mod threads;
pub use device_key::DeviceKeyBindingRecord;
pub use goals::ThreadGoalAccountingMode;
pub use goals::ThreadGoalAccountingOutcome;
pub use goals::ThreadGoalUpdate;
pub use remote_control::RemoteControlEnrollmentRecord;
pub use threads::ThreadFilterOptions;
File diff suppressed because it is too large Load Diff