mirror of
https://github.com/stablyai/orca.git
synced 2026-10-07 16:02:29 +00:00
* feat(agents): add first-class ZCode harness Add ZCode (Z.ai's `zcode` CLI) as a supervised Orca agent: managed lifecycle hooks on local, SSH and Windows hosts; status, question and approval reporting; synthetic status titles; session resume; orchestration worker launch options; and desktop + mobile agent-picker registration. Written against the newly open-sourced `zai-org/ZCode` (agent CLI 0.16.9), not against a remembered screen: - ZCode's hook runner writes a Claude-compatible stdin alias set, so it routes through the existing Claude-compatible vendor path while keeping its own identity in the sidebar. - `PermissionRequest` fires only once the approval card is on screen and racing the user's answer, so it is proof the pane is blocked, not an auto-approval. - ZCode's clarification tool is literally `AskUserQuestion` with Claude's questions/options shape, so Orca's question card renders it unchanged. - ZCode's `hooks.enabled` defaults to false, which is why configured hooks were reported as never firing; the installer sets it. - ZCode renames its own process to `zcode-cli`, so the expected foreground process cannot be the launch command or dispatch refuses the pane. - ZCode emits no OSC title in any state and repaints its ASCII banner forever, so readiness comes from Orca's synthetic hook title and launch drafts wait on the composer box rather than on a quiet render window. Three files crossed their max-lines limit, so each is split along a real seam: command-line entrypoint parsing out of agent process recognition, skill classification out of skill root discovery, and registry coverage out of the remote hook installer tests. Refs #10564 * fix(zcode): drop the session-option catalog and pin the orchestration contract ZCode's CLI exposes no `--model` flag at all, and the session-option launch path refuses to apply any option until a model id is chosen. A catalog therefore could not deliver `--mode` per worker, and would have accepted `--model` only to drop it silently. Take opencode's position instead: no catalog, so `worker-start --model` is refused with a clear message and ZCode launches with the model from its own config. `--mode` stays reachable through agent args, which is also how the yolo default is applied. Add a contract test covering the parts that make ZCode a usable worker: dispatchable foreground process, stdin prompt delivery, the prompt staying out of the launch command, and the composer-gated draft paste. * refactor(zcode): reuse shared helpers and cut the harness down No behaviour change; every ZCode test still passes. - Use installer-utils' own `hookDefinitionHasManagedCommand` instead of re-walking a hook definition by hand, which also drops a local string reader. - Share one `readZCodeEventMap` instead of keeping the same narrowing in both hook-settings and hook-config-json. - Collapse five identical error returns into one `zcodeHookError` builder, and return early from the status branches instead of assigning through `let`. - Split the event-to-status decision out of `normalizeZCodeEvent` into a pure `readZCodeTurn`, so the normalizer reads as decide-then-build and stops computing the tool name for events that never look at it. - Take a script file name in `readManagedZCodeHookEvents` like its siblings, which removes a `Parameters<typeof …>` indirection at the call site. - Drop the unused `ZCodeHookEvent` export and inline a single-use path helper. - Correct a stale comment: ZCode's loader is a strict `JSON.parse`, so the in-place edit preserves key order and indentation, not comments. * fix(zcode): address review — keep unmanaged event keys, correct comment, de-dupe README - `removeZCodeManagedHooks` deleted any event key whose list ended up empty, so an unrelated `"Notification": []` the user wrote was removed as collateral whenever a managed hook elsewhere made the write happen. Only touch an event Orca actually owned something in; covered by a new regression test. - The `isNewTurnEvent` comment claimed UserPromptSubmit was ZCode's only turn boundary while the expression below it also returned true for SessionStart. Say what the code does: SessionStart lands the idle boundary, UserPromptSubmit is the turn boundary (the Codex/Claude shape). - ZCode appeared twice in the README's single agent-badge block; keep the local-icon entry the link checker validates and drop the favicon duplicate. * docs(zcode): call out that the desktop bundle's CLI cannot open a session From live testing on #22464: pointing `zcode` at the desktop app's bundled `glm/zcode.cjs` installs Orca's hooks fine but then fails with `Cannot find package '@zcode/tui'`, so the pane never opens a session. The symptom reads as a broken harness when the CLI simply has no TUI. Say which build to use and how to check before reporting a problem. Reported-by: JWu527
196 lines
6.6 KiB
TypeScript
196 lines
6.6 KiB
TypeScript
import { homedir } from 'node:os'
|
|
import { join } from 'node:path'
|
|
import {
|
|
buildManagedCommandHook,
|
|
createManagedCommandMatcher,
|
|
getSharedManagedScriptPath,
|
|
hookDefinitionHasManagedCommand,
|
|
isPlainObject,
|
|
removeManagedCommands,
|
|
wrapPosixHookCommand,
|
|
wrapWindowsCmdHookCommand,
|
|
type HookDefinition
|
|
} from '../agent-hooks/installer-utils'
|
|
|
|
const ZCODE_SCRIPT_BASE = 'zcode-hook'
|
|
|
|
/**
|
|
* Every lifecycle event ZCode's hook runner can fire (`HookEventName` in
|
|
* `packages/contracts/src/hooks/index.ts`). Matchers are omitted on purpose:
|
|
* ZCode's `matchesAnyHookMatcher` treats an absent matcher as "every tool",
|
|
* and Claude's `"*"` is not a valid ZCode matcher.
|
|
*/
|
|
export const ZCODE_HOOK_EVENTS = [
|
|
'SessionStart',
|
|
'UserPromptSubmit',
|
|
'PreToolUse',
|
|
'PermissionRequest',
|
|
'PostToolUse',
|
|
'PostToolUseFailure',
|
|
'Stop'
|
|
] as const
|
|
|
|
/**
|
|
* ZCode's hook block, nested one level deeper than Claude's (`hooks.events.<Event>`).
|
|
*
|
|
* `events` is deliberately `unknown`-valued: it comes straight off a user-editable JSON
|
|
* file, so each entry is narrowed by `readEventDefinitions` at the point of use rather
|
|
* than asserted to be well-formed here.
|
|
*/
|
|
export type ZCodeHooksRuntimeConfig = {
|
|
enabled?: boolean
|
|
events?: Record<string, unknown>
|
|
[key: string]: unknown
|
|
}
|
|
|
|
export type ZCodeConfig = {
|
|
hooks?: ZCodeHooksRuntimeConfig
|
|
[key: string]: unknown
|
|
}
|
|
|
|
export function getZCodeConfigPath(): string {
|
|
// Why: ZCode resolves `~/.zcode/cli` from `homedir()` on every platform
|
|
// (`packages/adapters/src/config/file-config.adapter.ts`) — no APPDATA/XDG branch.
|
|
return join(homedir(), '.zcode', 'cli', 'config.json')
|
|
}
|
|
|
|
export function getZCodeRemoteConfigPath(remoteHome: string): string {
|
|
return `${remoteHome.replace(/\/$/, '')}/.zcode/cli/config.json`
|
|
}
|
|
|
|
export function getZCodeManagedScriptFileName(): string {
|
|
return process.platform === 'win32' ? `${ZCODE_SCRIPT_BASE}.cmd` : `${ZCODE_SCRIPT_BASE}.sh`
|
|
}
|
|
|
|
export function getZCodePosixManagedScriptFileName(): string {
|
|
return `${ZCODE_SCRIPT_BASE}.sh`
|
|
}
|
|
|
|
export function getZCodeManagedScriptPath(): string {
|
|
return getSharedManagedScriptPath(getZCodeManagedScriptFileName())
|
|
}
|
|
|
|
export function getZCodeManagedCommand(scriptPath: string): string {
|
|
if (process.platform === 'win32') {
|
|
// Why: ZCode spawns a `type: "command"` hook through its own shell resolver, so keep the
|
|
// bare directly-spawnable .cmd on the safe path and fall back to the encoded form otherwise.
|
|
return wrapWindowsCmdHookCommand(scriptPath)
|
|
}
|
|
return wrapPosixHookCommand(scriptPath)
|
|
}
|
|
|
|
export function getZCodeRemoteManagedCommand(scriptPath: string): string {
|
|
return wrapPosixHookCommand(scriptPath)
|
|
}
|
|
|
|
function getZCodeManagedCommandMatcher(
|
|
scriptFileName = getZCodeManagedScriptFileName()
|
|
): (command: string | undefined) => boolean {
|
|
return createManagedCommandMatcher(scriptFileName)
|
|
}
|
|
|
|
/** The `hooks.events` block as a plain lookup, or empty when absent or malformed. */
|
|
export function readZCodeEventMap(config: ZCodeConfig): Record<string, unknown> {
|
|
const events = config.hooks?.events
|
|
return isPlainObject(events) ? events : {}
|
|
}
|
|
|
|
/** The definitions registered for one event, dropping anything not shaped like a list. */
|
|
function readEventDefinitions(
|
|
events: Record<string, unknown>,
|
|
eventName: string
|
|
): HookDefinition[] {
|
|
const definitions = events[eventName]
|
|
if (!Array.isArray(definitions)) {
|
|
return []
|
|
}
|
|
// Why: a hand-edited config can hold nulls or scalars here; keep only object entries so
|
|
// the callers below never have to re-check, and never throw on user content.
|
|
return definitions.filter((definition): definition is HookDefinition => isPlainObject(definition))
|
|
}
|
|
|
|
export function applyZCodeManagedHooks(
|
|
config: ZCodeConfig,
|
|
command: string,
|
|
scriptFileName = getZCodeManagedScriptFileName()
|
|
): ZCodeConfig {
|
|
const nextEvents = { ...readZCodeEventMap(config) }
|
|
const isManagedCommand = getZCodeManagedCommandMatcher(scriptFileName)
|
|
|
|
for (const eventName of ZCODE_HOOK_EVENTS) {
|
|
const current = readEventDefinitions(nextEvents, eventName)
|
|
const cleaned = removeManagedCommands(current, isManagedCommand)
|
|
nextEvents[eventName] = [...cleaned, { hooks: [buildManagedCommandHook(command)] }]
|
|
}
|
|
|
|
return {
|
|
...config,
|
|
hooks: {
|
|
...config.hooks,
|
|
// Why: ZCode's DefaultRuntimeConfig ships `hooks.enabled: false`, so a hook block alone
|
|
// fires nothing — this flag is what the "ZCode hooks never run" reports were missing.
|
|
enabled: true,
|
|
events: nextEvents
|
|
}
|
|
}
|
|
}
|
|
|
|
export function removeZCodeManagedHooks(
|
|
config: ZCodeConfig,
|
|
scriptFileName = getZCodeManagedScriptFileName()
|
|
): { config: ZCodeConfig; changed: boolean } {
|
|
const events = readZCodeEventMap(config)
|
|
const nextEvents = { ...events }
|
|
const isManagedCommand = getZCodeManagedCommandMatcher(scriptFileName)
|
|
let changed = false
|
|
|
|
for (const [eventName, value] of Object.entries(nextEvents)) {
|
|
// Why: leave a non-array value exactly as the user wrote it — emptying it below would
|
|
// delete a key Orca never owned.
|
|
if (!Array.isArray(value)) {
|
|
continue
|
|
}
|
|
const definitions = readEventDefinitions(nextEvents, eventName)
|
|
const cleaned = removeManagedCommands(definitions, isManagedCommand)
|
|
// Why: only touch an event Orca actually owned something in. Without this, an unrelated
|
|
// empty entry the user wrote (`"Notification": []`) was deleted as collateral whenever a
|
|
// managed hook elsewhere made the write happen.
|
|
if (JSON.stringify(cleaned) === JSON.stringify(definitions)) {
|
|
continue
|
|
}
|
|
changed = true
|
|
if (cleaned.length === 0) {
|
|
delete nextEvents[eventName]
|
|
} else {
|
|
nextEvents[eventName] = cleaned
|
|
}
|
|
}
|
|
|
|
if (!changed) {
|
|
return { config, changed: false }
|
|
}
|
|
// Why: leave `hooks.enabled` alone on remove — the user may run their own hooks, and
|
|
// flipping it back to false would silently disable those too.
|
|
return { config: { ...config, hooks: { ...config.hooks, events: nextEvents } }, changed: true }
|
|
}
|
|
|
|
/** Events whose managed command is currently registered in the user's config. */
|
|
export function readManagedZCodeHookEvents(
|
|
config: ZCodeConfig,
|
|
scriptFileName = getZCodeManagedScriptFileName()
|
|
): Set<string> {
|
|
const isManagedCommand = getZCodeManagedCommandMatcher(scriptFileName)
|
|
const events = readZCodeEventMap(config)
|
|
return new Set(
|
|
ZCODE_HOOK_EVENTS.filter((eventName) =>
|
|
readEventDefinitions(events, eventName).some((definition) =>
|
|
hookDefinitionHasManagedCommand(definition, isManagedCommand)
|
|
)
|
|
)
|
|
)
|
|
}
|
|
|
|
export function isZCodeHooksEnabled(config: ZCodeConfig): boolean {
|
|
return config.hooks?.enabled === true
|
|
}
|