Files
orca/src/shared/structured-agent-session-composer.ts
T
Brennan Benson 563dd5487f feat(native-chat): show a Codex chat's goal above the composer, and set it from goal mode (#22377)
* feat(native-chat): show a Codex chat's goal above the composer and set it from goal mode

Structured Codex chat now treats the thread goal as session state: a banner above the
composer shows the current goal (pursuing / paused) with clear, pause/resume and expand;
/goal enters a goal mode whose send calls thread/goal/set; the objective is journaled as a
user message marked as sent as a goal. The banner is derived from the journaled goal rows,
which Codex's resume snapshot refreshes, so a reopened or adopted chat shows its goal.

Fixes STA-8159

* fix(native-chat): replace a recorded goal by clearing first, and recover a lost goal-change response

- A set while the journal records a goal (any status) clears it before setting,
  so the new goal starts with its own time and token counters instead of
  rewriting the old goal's objective in place.
- The threadGoal plan answers an unknown outcome from the goal the journal
  records and reruns otherwise, so one request timeout no longer refuses every
  later Clear/Pause/Resume as unknown for the mounted session.
- The goal-mode chip says "Exit goal mode"; "Clear goal" stays the banner's
  action on the provider goal.
- A typed bare /goal on Enter enters goal mode, the same as picking it.
- The renderer reads the goal off the tail of its ordered snapshot; the host's
  unordered map keeps the by-sequence reader.
- Drop the composer's duplicate in-flight guard; the goal controller already
  serializes changes.
- Pin that a counter-only revision reaches a subscriber's live page under its
  original sequence.

* fix(native-chat): keep a bare /goal inside goal mode as the entrance, and pin goal delivery and serialization

- A bare `/goal` submitted while already in goal mode re-enters the mode instead
  of setting a goal whose objective is the literal text "/goal".
- The counter-only revision pin now drives the host's own event sink bound to a
  real journal, so it goes red when the publish after a lifecycle transition is
  dropped; the previous fake sink never published.
- Pin that a set which threw after journaling its objective puts that objective
  back exactly once when the ledger reruns the same operation id.
- Cover the goal controller hook: absent without host support, the loaded window
  wins over the host's answer, a second change while one is unsettled answers
  false without a request, and a refused change frees the next one.

* fix(native-chat): resume a blocked or usage-limited goal, and keep goal-mode drafts honest

- The goal bar offers Resume on a blocked or usage-limited goal, which the
  provider resumes exactly as it resumes a paused one; a goal whose token budget
  is spent still offers only Clear. The rule lives beside the other goal facts
  in shared code so every reader answers it the same way.
- A `/goal <text>` typed inside goal mode sets the objective `<text>`, as it
  does outside goal mode, instead of a goal whose objective is the literal
  command.
- Setting a goal is a host round trip; a draft edited while it was in flight is
  no longer wiped when the goal lands, matching every other host command.
- Pin that a lost status-change response is read as applied only when the
  recorded goal is in that status, that a cleared row in the loaded window
  outranks the host's earlier answer, and that the PTY lane is untouched.

* fix(native-chat): keep the load-older anchor on the loaded window when a live revision lands below it

A live revision of a row keeps that row's original sequence. When the row is
older than the client's loaded window, the shared reducer merged it in and it
became the load-older anchor, so paging `before` it skipped every row between.
A goal's counter-only revisions during a long goal turn reach any client that
attached after the goal row left its window, so a reopened chat lost rows on
scroll-back.

The reducer now admits live rows only at or above the window's oldest row
while older rows remain on the host; the journal keeps the revision and the
page reader serves it once the window reaches the row. With nothing older on
the host the window is the whole journal, so a row below the head is admitted
as before.

Also drain accepted provider events before a goal set reads the journal to
decide whether it replaces a recorded goal.
2026-09-23 10:34:06 -07:00

206 lines
7.6 KiB
TypeScript

import {
getHostClaimedNativeChatCommands,
getTextDrivenNativeChatCommands
} from './native-chat-agent-profiles'
import type { AgentType } from './agent-status-types'
import type { SessionOptionDescriptor, SessionOptionValue } from './native-chat-session-options'
import type { SlashCommandSuggestion } from './native-chat-slash-commands'
import type { AgentSessionConversationCommand } from './agent-session-conversation-command'
const MODEL_COMMAND: SlashCommandSuggestion = {
name: 'model',
description: 'Choose the model'
}
const EFFORT_COMMAND: SlashCommandSuggestion = {
name: 'effort',
description: 'Choose reasoning effort'
}
const CONVERSATION_COMMANDS: readonly SlashCommandSuggestion[] = [
{ name: 'clear', description: 'Start a fresh conversation' },
{ name: 'compact', description: 'Compact conversation context' }
]
/** Session options remain available on hosts predating conversation commands. */
export const STRUCTURED_AGENT_SESSION_SLASH_COMMANDS: readonly SlashCommandSuggestion[] = [
MODEL_COMMAND,
EFFORT_COMMAND
]
export type StructuredAgentSessionComposerOptions = {
agent?: AgentType
snapshot: readonly SessionOptionDescriptor[]
invokeAction: (id: string) => Promise<boolean>
setOption: (id: string, value: SessionOptionValue) => Promise<boolean>
conversationCommands?: readonly AgentSessionConversationCommand[]
runConversationCommand?: (
command: AgentSessionConversationCommand
) => Promise<{ accepted: boolean; error: string | null }>
/** Present only where the host can set this session's goal; otherwise `/goal`
* stays message text the agent acts on itself. */
setThreadGoalObjective?: (objective: string) => Promise<boolean>
}
export type StructuredAgentSessionCommandOutcome = {
handled: boolean
accepted: boolean
error: string | null
}
function commandParts(text: string): { name: string; argument: string } | null {
if (!text.startsWith('/')) {
return null
}
const match = /^\/([^\s]+)(?:\s+(.*))?$/.exec(text.trimEnd())
return match ? { name: match[1]!.toLowerCase(), argument: match[2]?.trim() ?? '' } : null
}
/** The commands the composer menu offers when the host reports no catalog of its
* own: the host's own commands, plus the ones this agent acts on from message
* text. Both are honored — the first here, the second by the agent — so a menu
* pick is never answered with "not available". */
export function structuredSlashCommands(
commands: readonly AgentSessionConversationCommand[] = [],
agent?: AgentType | null
): readonly SlashCommandSuggestion[] {
const hostOwned = [
...STRUCTURED_AGENT_SESSION_SLASH_COMMANDS,
...CONVERSATION_COMMANDS.filter((entry) =>
commands.includes(entry.name as AgentSessionConversationCommand)
)
]
// Why: a host with no catalog to report would otherwise hide the commands the
// agent itself implements, e.g. Codex's `/goal`.
return [
...hostOwned,
...getTextDrivenNativeChatCommands(agent).filter(
(entry) => !hostOwned.some((offered) => offered.name === entry.name)
)
]
}
/** Wider than the offered menu on purpose: a TUI-only command still has to be
* claimed here and answered, or a hand-typed `/clear` reaches the model as
* literal prompt text. Commands the agent itself implements are deliberately
* absent — the profile unclaims those so they pass through as text. */
function structuredRecognizedCommands(agent: AgentType): readonly SlashCommandSuggestion[] {
return [
...STRUCTURED_AGENT_SESSION_SLASH_COMMANDS,
...CONVERSATION_COMMANDS,
...getHostClaimedNativeChatCommands(agent)
]
}
/** Whether the chat host, rather than the agent, owns this command. Callers also
* use it to refuse attachments: a host command sends no message, so attachments
* would be silently dropped, whereas a pass-through command is a real send. */
export function isStructuredAgentSessionComposerCommand(
text: string,
agent: AgentType = 'codex'
): boolean {
const command = commandParts(text)
return Boolean(
command && structuredRecognizedCommands(agent).some((entry) => entry.name === command.name)
)
}
/** `/goal …`, which the host answers only where it can set this session's goal. */
export function isStructuredAgentSessionGoalCommand(text: string): boolean {
return commandParts(text)?.name === 'goal'
}
/** `/goal` with nothing after it: an entrance to goal mode, not an objective. */
export function isBareStructuredAgentSessionGoalCommand(text: string): boolean {
const command = commandParts(text)
return command?.name === 'goal' && command.argument === ''
}
/** The objective a goal-mode draft names. A `/goal …` typed there out of habit
* names the same objective it would outside goal mode, never the literal command. */
export function structuredAgentSessionGoalObjective(text: string): string {
const command = commandParts(text)
return command?.name === 'goal' ? command.argument : text.trim()
}
function unavailable(name: string): StructuredAgentSessionCommandOutcome {
return {
handled: true,
accepted: true,
error: `/${name} is not available in chat sessions. Use the slash menu to see available commands.`
}
}
export async function dispatchStructuredAgentSessionComposerCommand(
text: string,
controller: StructuredAgentSessionComposerOptions
): Promise<StructuredAgentSessionCommandOutcome> {
const command = commandParts(text)
if (command?.name === 'goal' && controller.setThreadGoalObjective) {
if (!command.argument) {
return { handled: true, accepted: false, error: 'Describe the goal after /goal.' }
}
// A refusal reaches the user through the session's own error surface.
return {
handled: true,
accepted: await controller.setThreadGoalObjective(command.argument),
error: null
}
}
if (!command || !isStructuredAgentSessionComposerCommand(text, controller.agent)) {
return { handled: false, accepted: false, error: null }
}
if (command.name === 'clear' || command.name === 'compact') {
if (command.argument) {
return { handled: true, accepted: false, error: `Use /${command.name} without arguments.` }
}
if (
!controller.conversationCommands?.includes(command.name) ||
!controller.runConversationCommand
) {
return {
handled: true,
accepted: false,
error: `/${command.name} is not supported by this chat host.`
}
}
return { handled: true, ...(await controller.runConversationCommand(command.name)) }
}
if (command.name !== 'model' && command.name !== 'effort') {
return unavailable(command.name)
}
const descriptor = controller.snapshot.find((entry) => entry.id === command.name)
if (!descriptor || descriptor.kind.type !== 'select') {
return {
handled: true,
accepted: true,
error: `${command.name === 'model' ? 'Models' : 'Reasoning effort'} are unavailable for this chat session.`
}
}
if (!command.argument) {
const opened = await controller.invokeAction(command.name)
return {
handled: true,
accepted: opened,
error: opened ? null : `Could not open the ${command.name} picker.`
}
}
const normalized = command.argument.toLowerCase()
const choice = descriptor.kind.choices.find(
(entry) => entry.value.toLowerCase() === normalized || entry.label.toLowerCase() === normalized
)
if (!choice) {
return {
handled: true,
accepted: false,
error: `${command.argument} is not an available ${command.name} for this chat session.`
}
}
const applied = await controller.setOption(command.name, choice.value)
return {
handled: true,
accepted: applied,
error: applied ? null : `Could not apply ${command.name} ${choice.label}.`
}
}