Files
Michael Ramos 82a8f236ec feat(opencode): restore the slash commands on OpenCode 2 (#1434)
* feat(opencode): restore the slash commands on OpenCode 2

OpenCode's V2 plugin API gained native command execution upstream
(anomalyco/opencode issue #2185, PR #44765): ctx.command.transform lets a
plugin add a command whose execute callback fully owns the invocation. That
shape currently ships on the beta and dev dist-tags of @opencode-ai/plugin
while next and latest still carry the older context, so the capability is
duck-typed at runtime and never imported. On a host that exposes it the V2
adapter registers /plannotator-review, /plannotator-annotate and
/plannotator-last and runs the same handleCliCommand machinery OpenCode 1
uses, passing the raw argument tail straight through to the CLI. On a host
without it nothing new is registered and behavior is byte-identical to before.

Also wires ctx.session.switchAgent (same API generation, same probe) so an
agent switch chosen in the review UI is applied instead of only warned about,
and accepts both agent.list() response shapes: the HTTP client types it as a
{ location, data } envelope while the in-process plugin domain answers with a
bare array, where reading .data threw and silently emptied the agent list.

The shared command stubs get model-mediated fallback bodies for OpenCode 2
hosts on the stale channels. They carry no shell interpolation on purpose:
OpenCode 1 evaluates a template's !`...` before the V1 plugin's
command.execute.before hook can clear the parts, so a bang template there
would launch a second Plannotator session on every OC1 invocation. A source
level test pins that.

AI-assisted (Claude) under maintainer direction.

* fix(opencode): probe the command draft and reclaim the names from the stubs

Review found the capability probe was wrong in the direction that matters.
ctx.command.transform exists on pre-#44765 hosts too: our own pinned
@opencode-ai/plugin@0.0.0-next-16775 declares CommandDraft as
{ list, get, update, remove } with no add. The probe therefore returned true on
next and latest, draft.add was undefined, and because transforms are stored and
replayed the TypeError landed in the batched reload flush and aborted it before
commit, plausibly taking every command registration on the host down with it.
Capability is now read from the draft handed to the callback, which is the only
witness, and the registration call is wrapped so no transform rejection can fail
plugin setup.

The stubs also shadowed the native definitions on new hosts. Command definitions
land in a name-keyed map where add is Map.set, transforms replay in registration
order, and OpenCode's own ConfigCommandPlugin activates in the post group after
package plugins while scanning the exact directory the installer writes the
three stubs to. A setup-time registration is therefore always overwritten on a
normal install. The plugin now re-registers the same transform once activation
settles, so its definitions are last in the replay order, and calls
ctx.command.reload() explicitly because a late registration only adds its reload
to the already-flushed boot batch. Ownership is read back from
ctx.command.list() by description, which is why the native descriptions and the
stub frontmatter are deliberately distinct. If the reclaim cannot run the stubs
keep the names and the commands still work through their fallback bodies.

Also: a failing switchAgent no longer costs the reviewer their feedback on the
command path, feedback is delivered as "queue" rather than replaying the
invocation's admission mode minutes later when a steer would land mid-turn, and
the agent-list comment no longer asserts a bare-array response that could not be
reproduced upstream (accepting both shapes is still right, since reading .data
blindly throws into a catch that degrades silently).

Tests: the real old-host draft shape registers nothing and throws nothing, the
shadowing contest is modelled against upstream's replay semantics, the OpenCode 1
parts-clearing invariant is pinned for all three commands in both plan-agent and
manual mode now that the stubs carry real instructions, and the V2 smoke asserts
the plugin did not activate as failed and that all three commands resolve. The
smoke now also installs the stubs into its sandbox config dir so the contest
actually happens there. scripts/opencode2-native-commands-smoke.sh runs the same
smoke against a dev-channel build with native commands required; CI cannot,
because it pins a next build.

AI-assisted (Claude) under maintainer direction.

* fix(opencode): keep the reclaim ticking and stop an unbuilt checkout failing setup

The reclaim ended the loop when the draft-probe flag read false, but that flag
only flips when the transform replays, which under boot batching is the flush
after every plugin has loaded. Plannotator loads before the post-group config
plugins, so the first tick legitimately reads false and the loop exited for
good: the reclaim was inert in exactly the shape production has. The tick is
skipped now instead, with a test that flips the flag between ticks.

The V1 entry called resolveBundledHtmlPath synchronously during plugin
construction, outside the .catch that was there to absorb a missing asset, so an
unbuilt checkout threw out of construction before any code path that needs the
HTML. The Test workflow runs bun test with no build step, so the new OpenCode 1
interception tests failed there. Both preloads are guarded; the lazy getters
still raise a clear error if something actually needs the file.

The smoke's failed-plugin guard read entry.state.status, but Plugin.Info carries
status and error at the top level, so a failed activation slipped through.
Reads the top level first and keeps the nested one as a fallback.

Comment corrections: State.batch clears its active flag before flushing, so a
late transform registration materializes on its own; the explicit reload() is
redundant-but-defensive rather than required. The reclaim schedule is a list of
deltas the loop awaits in turn, so the ticks land near 0.3s, 1.5s, 5.5s and
15.5s, not at the raw numbers.

AI-assisted (Claude) under maintainer direction.
2026-08-31 10:42:26 -07:00

602 lines
22 KiB
TypeScript

/**
* Plannotator Plugin for OpenCode
*
* POC: Edit-based submit_plan. The tool accepts line-range edits instead of
* full plan text or file paths. A backing file is managed by the plugin;
* the agent never touches it directly. On denial, the tool response includes
* the plan with line numbers so the agent can target surgical edits.
*
* Environment variables:
* PLANNOTATOR_REMOTE - Set to "1"/"true" for remote, "0"/"false" for local
* PLANNOTATOR_PORT - Fixed port to use (default: random locally, 19432 for remote)
* PLANNOTATOR_PLAN_TIMEOUT_SECONDS - Max wait for approval (default: 345600, set 0 to disable)
* PLANNOTATOR_ALLOW_SUBAGENTS - Set to "1" to allow subagents to see submit_plan
*
* @packageDocumentation
*/
import { type Plugin, tool } from "@opencode-ai/plugin/v1";
import { existsSync, readFileSync } from "node:fs";
import { readFile } from "node:fs/promises";
import path from "node:path";
import { fileURLToPath, pathToFileURL } from "node:url";
import { loadConfig, resolveSharingEnabled } from "@plannotator/shared/config";
import { readImprovementHook } from "@plannotator/shared/improvement-hooks";
import { composeImproveContext } from "@plannotator/shared/pfm-reminder";
import {
composeSystemPrompt,
stripConflictingPlanModeRules,
} from "./plan-mode";
import {
applyWorkflowConfig,
isPlanningAgent,
normalizeWorkflowOptions,
shouldApplyToolDefinitionRewrites,
shouldInjectFullPlanningPrompt,
shouldInjectGenericPlanReminder,
shouldModifyPrompts,
shouldRegisterSubmitPlan,
type RuntimeMode,
type PlannotatorOpenCodeOptions,
} from "./workflow";
import {
handleCliCommand,
runCliPlanReview,
type OpenCodeBridgeContext,
type OpenCodePlanReviewResult,
} from "./cli-bridge";
import { resolveValidatedTargetAgent } from "./agent-switch";
import { shouldFallbackAfterEmbeddedError } from "./prompt-delivery-error";
import { executeSubmitPlan } from "./submit-plan-executor";
import { getPlanningPrompt } from "./planning-prompt";
// Lazy-load HTML at first use instead of embedding in the bundle.
// The two SPA files are ~20 MB combined — inlining them as string literals
// adds ~160ms to module parse time (see GitHub issue #410).
let _planHtml: string | null = null;
let _reviewHtml: string | null = null;
const moduleDir = path.dirname(fileURLToPath(import.meta.url));
function resolveBundledHtmlPath(filename: string): string {
const candidates = [
path.join(moduleDir, filename),
path.join(moduleDir, "..", filename),
];
for (const candidate of candidates) {
if (existsSync(candidate)) {
return candidate;
}
}
throw new Error(`Could not find bundled HTML asset: ${filename}`);
}
function readBundledHtml(filename: string): string {
return readFileSync(resolveBundledHtmlPath(filename), "utf-8");
}
/** Best-effort warm of the sync cache. Never throws, on any failure. */
function preloadBundledHtml(filename: string, assign: (html: string) => void): void {
try {
readFile(resolveBundledHtmlPath(filename), "utf-8").then(assign).catch(() => {});
} catch {
// The asset is not on disk. The lazy getters raise a clear error if and
// when a code path actually needs it.
}
}
function getPlanHtml(): string {
if (!_planHtml) _planHtml = readBundledHtml("plannotator.html");
return _planHtml;
}
function getReviewHtml(): string {
if (!_reviewHtml) _reviewHtml = readBundledHtml("review-editor.html");
return _reviewHtml;
}
const DEFAULT_PLAN_TIMEOUT_SECONDS = 345_600; // 96 hours
// ── Plugin ────────────────────────────────────────────────────────────────
function getLastUserAgentFromMessages(messages: any[] | undefined): string | undefined {
if (!messages) return undefined;
for (let i = messages.length - 1; i >= 0; i--) {
const msg = messages[i];
if (msg.info?.role === "user" && typeof msg.info.agent === "string") {
return msg.info.agent;
}
}
return undefined;
}
function getBunRuntime(): { serve?: unknown; sleep?: (ms: number) => Promise<void> } | undefined {
return (globalThis as typeof globalThis & {
Bun?: { serve?: unknown; sleep?: (ms: number) => Promise<void> };
}).Bun;
}
function hasEmbeddedRuntime(): boolean {
return typeof getBunRuntime()?.serve === "function";
}
function shouldUseEmbeddedRuntime(runtime: RuntimeMode): boolean {
return runtime !== "cli" && hasEmbeddedRuntime();
}
function getEmbeddedRuntimeError(): string {
return "runtime \"embedded\" requires a Bun-hosted OpenCode plugin runtime. Use runtime \"auto\" or \"cli\" with this OpenCode host.";
}
function logPlannotatorReady(client: any, label: string, url: string): void {
try {
void client.app.log({ level: "info", message: `[Plannotator] Open ${label}: ${url}` });
} catch {
// OpenCode logging is best-effort.
}
// app.log only reaches OpenCode's server log file — never the TUI. Toast the
// URL too so remote users (no auto-opened browser) actually see it.
// Best-effort: older hosts without /tui/show-toast just no-op.
try {
const result = client.tui?.showToast?.({
body: { title: "Plannotator", message: `Open ${label}: ${url}`, variant: "info" },
});
// A fetch-level failure (host restarting) rejects the SDK promise; swallow
// it so a cosmetic toast can never surface an unhandled rejection.
if (result && typeof result.catch === "function") result.catch(() => {});
} catch {
// Toast delivery is best-effort.
}
}
type EmbeddedRuntimeModule = {
runEmbeddedPlanReview: (input: {
client: any;
planContent: string;
sharingEnabled: boolean;
shareBaseUrl?: string;
pasteApiUrl?: string;
htmlContent: string;
timeoutSeconds: number | null;
abortSignal: AbortSignal;
logReady: (url: string, isRemote: boolean, port: number) => void;
}) => Promise<OpenCodePlanReviewResult>;
handleEmbeddedCommand: (
command: string,
event: any,
deps: {
client: any;
htmlContent: string;
reviewHtmlContent: string;
getSharingEnabled: () => Promise<boolean>;
getShareBaseUrl: () => string | undefined;
getPasteApiUrl: () => string | undefined;
directory?: string;
},
) => Promise<{ feedback?: string | null }>;
};
async function importEmbeddedRuntime(): Promise<EmbeddedRuntimeModule> {
const builtPath = path.join(moduleDir, "embedded.js");
if (existsSync(builtPath)) {
return await import(pathToFileURL(builtPath).href) as EmbeddedRuntimeModule;
}
const sourceSpecifier = "./embedded";
return await import(sourceSpecifier) as EmbeddedRuntimeModule;
}
/** Run one OpenCode plan review in the configured runtime. */
async function runPlanReview(input: {
client: any;
runtime: RuntimeMode;
planContent: string;
sharingEnabled: boolean;
shareBaseUrl?: string;
pasteApiUrl?: string;
htmlContent: string;
timeoutSeconds: number | null;
abortSignal: AbortSignal;
cwd?: string;
bridge: OpenCodeBridgeContext;
}): Promise<OpenCodePlanReviewResult> {
input.abortSignal.throwIfAborted();
if (input.runtime === "embedded" && !hasEmbeddedRuntime()) {
throw new Error(getEmbeddedRuntimeError());
}
if (shouldUseEmbeddedRuntime(input.runtime)) {
try {
const embedded = await importEmbeddedRuntime();
return await embedded.runEmbeddedPlanReview({
client: input.client,
planContent: input.planContent,
sharingEnabled: input.sharingEnabled,
shareBaseUrl: input.shareBaseUrl,
pasteApiUrl: input.pasteApiUrl,
htmlContent: input.htmlContent,
timeoutSeconds: input.timeoutSeconds,
abortSignal: input.abortSignal,
logReady: (url) => logPlannotatorReady(input.client, "plan review", url),
});
} catch (error) {
input.abortSignal.throwIfAborted();
if (input.runtime === "embedded") throw error;
try {
void input.client.app.log({
level: "error",
message: `[Plannotator] Embedded runtime unavailable; falling back to CLI: ${error instanceof Error ? error.message : String(error)}`,
});
} catch {}
}
}
return await runCliPlanReview({
client: input.client,
planContent: input.planContent,
cwd: input.cwd,
timeoutSeconds: input.timeoutSeconds,
abortSignal: input.abortSignal,
bridge: input.bridge,
});
}
const PlannotatorPlugin: Plugin = async (ctx, rawOptions?: PlannotatorOpenCodeOptions) => {
const workflowOptions = normalizeWorkflowOptions(rawOptions);
// Preload HTML in background: populates the sync cache before first use.
// `resolveBundledHtmlPath` THROWS when the asset is absent, and it runs
// synchronously here, outside the .catch that was meant to absorb exactly
// that. An unbuilt checkout (or a partial install) therefore took down plugin
// construction itself, before any code path that needs the HTML. A missing
// asset must only fail the feature that reads it, which is what the lazy
// getters already do.
preloadBundledHtml("plannotator.html", (html) => { _planHtml = html; });
preloadBundledHtml("review-editor.html", (html) => { _reviewHtml = html; });
let cachedAgents: any[] | null = null;
async function getSharingEnabled(): Promise<boolean> {
try {
const response = await ctx.client.config.get({ query: { directory: ctx.directory } });
// @ts-ignore - share config may exist
const share = response?.data?.share;
if (share !== undefined) {
return share !== "disabled";
}
} catch {
// Config read failed, fall through to env var / plannotator config
}
return resolveSharingEnabled(loadConfig());
}
function getShareBaseUrl(): string | undefined {
return process.env.PLANNOTATOR_SHARE_URL || undefined;
}
function getPasteApiUrl(): string | undefined {
return process.env.PLANNOTATOR_PASTE_URL || undefined;
}
async function getOpenCodeAgents(): Promise<any[] | undefined> {
try {
if (!cachedAgents) {
const response = await ctx.client.app.agents({
query: { directory: ctx.directory },
});
cachedAgents = response.data ?? [];
}
return cachedAgents;
} catch {
return undefined;
}
}
async function getBridgeContext(): Promise<OpenCodeBridgeContext> {
return {
sharingEnabled: await getSharingEnabled(),
shareBaseUrl: getShareBaseUrl(),
pasteApiUrl: getPasteApiUrl(),
agents: await getOpenCodeAgents(),
};
}
function getPlanTimeoutSeconds(): number | null {
const raw = process.env.PLANNOTATOR_PLAN_TIMEOUT_SECONDS?.trim();
if (!raw) return DEFAULT_PLAN_TIMEOUT_SECONDS;
const parsed = Number.parseInt(raw, 10);
if (!Number.isFinite(parsed) || parsed < 0) {
console.error(
`[Plannotator] Invalid PLANNOTATOR_PLAN_TIMEOUT_SECONDS="${raw}". Using default ${DEFAULT_PLAN_TIMEOUT_SECONDS}s.`
);
return DEFAULT_PLAN_TIMEOUT_SECONDS;
}
if (parsed === 0) return null;
return parsed;
}
function allowSubagents(): boolean {
const val = process.env.PLANNOTATOR_ALLOW_SUBAGENTS?.trim();
return val === "1" || val === "true";
}
const plugin: any = {
config: async (opencodeConfig) => {
applyWorkflowConfig(opencodeConfig, workflowOptions, allowSubagents());
},
// Replace OpenCode's "STRICTLY FORBIDDEN" plan mode prompt with a version
// that allows markdown file writing. OpenCode's original blocks ALL file edits,
// but we need the agent to write plans, specs, docs, etc.
"experimental.chat.messages.transform": async (input, output) => {
if (!shouldModifyPrompts(workflowOptions)) return;
const lastUserAgent = getLastUserAgentFromMessages(output.messages);
if (
workflowOptions.workflow === "plan-agent"
&& !isPlanningAgent(lastUserAgent, workflowOptions)
) {
return;
}
for (const message of output.messages) {
if (message.info.role !== "user") continue;
for (const part of message.parts as any[]) {
if (part.type !== "text" || !part.text?.includes("STRICTLY FORBIDDEN")) continue;
part.text = `<system-reminder>
# Plan Mode - System Reminder
CRITICAL: Plan mode ACTIVE. You are in a PLANNING phase. The ONLY file modifications
allowed are writing or editing markdown files (.md) — plans, specs, documentation, etc.
All other file edits, code modifications, and system changes are STRICTLY FORBIDDEN.
Do NOT use bash commands to manipulate non-markdown files. Commands may ONLY read/inspect.
## Responsibility
Your responsibility is to think, read, search, and delegate explore agents to construct
a well-formed plan. Ask the user clarifying questions and surface tradeoffs rather than
making assumptions about intent. Use submit_plan to submit your plan for user review.
## Important
The user wants a plan, not execution. You MUST NOT edit source code, run non-readonly
tools (except writing markdown files), or otherwise make changes to the system.
</system-reminder>`;
}
}
},
// Suppress plan_exit — redirect to submit_plan
// Override todowrite — defer to submit_plan during planning
"tool.definition": async (input, output) => {
if (!shouldApplyToolDefinitionRewrites(workflowOptions)) return;
if (input.toolID === "plan_exit") {
output.description =
"Do not call this tool. Use submit_plan instead — it opens a visual review UI for plan approval.";
}
if (input.toolID === "todowrite") {
output.description =
"While actively planning with the user, use submit_plan instead. Only use todos once implementation begins or unless the user explicitly asks.";
}
},
// Inject planning instructions into system prompt
"experimental.chat.system.transform": async (input, output) => {
if (!shouldModifyPrompts(workflowOptions)) return;
const systemText = output.system.join("\n");
if (systemText.toLowerCase().includes("title generator") || systemText.toLowerCase().includes("generate a title")) {
return;
}
let lastUserAgent: string | undefined;
let isSubagent = false;
try {
const messagesResponse = await ctx.client.session.messages({
// @ts-ignore - sessionID exists on input
path: { id: input.sessionID }
});
const messages = messagesResponse.data;
lastUserAgent = getLastUserAgentFromMessages(messages);
if (!lastUserAgent) return;
// Cache agents list (static per session)
if (!cachedAgents) {
const agentsResponse = await ctx.client.app.agents({
query: { directory: ctx.directory }
});
cachedAgents = agentsResponse.data ?? [];
}
const agent = cachedAgents.find((a: { name: string }) => a.name === lastUserAgent);
// @ts-ignore - Agent has mode field
isSubagent = agent?.mode === "subagent";
} catch {
return;
}
if (shouldInjectFullPlanningPrompt(lastUserAgent, workflowOptions)) {
const stripped = stripConflictingPlanModeRules(output.system);
const hook = readImprovementHook("enterplanmode-improve");
const pfmEnabled = loadConfig().pfmReminder === true;
const improveContext = composeImproveContext({
pfmEnabled,
improvementHookContent: hook?.content ?? null,
});
const parts = [...stripped, getPlanningPrompt()];
if (improveContext) parts.push(improveContext);
output.system.length = 0;
output.system.push(...composeSystemPrompt([], parts));
return;
}
if (!shouldInjectGenericPlanReminder(lastUserAgent, isSubagent, workflowOptions)) return;
const planSubmissionReminder = `## Plan Submission
When you have completed your plan, call the \`submit_plan\` tool to submit it for user review. Pass your full plan as a single edit: \`{ "edits": [{ "start": 1, "content": "..." }] }\`.
The user will review your plan in a visual UI where they can annotate, approve, or request changes. If rejected, the response includes your plan with line numbers; use targeted edits to revise specific sections.
Do NOT proceed with implementation until your plan is approved.`;
const composed = composeSystemPrompt(output.system, [planSubmissionReminder]);
output.system.length = 0;
output.system.push(...composed);
},
// Intercept plannotator commands before the agent sees them.
// Clearing output.parts in place suppresses the .md body + appended
// args so the agent never receives the command — without this, OpenCode
// calls resolvePromptParts() on "<body> <arguments>", which auto-attaches
// any file path it finds as a FilePart. On a large file that blows the
// context before the annotation UI even opens (#713).
//
// Must mutate in place (length = 0), not reassign (= []). The caller
// holds a reference to the parts array directly and ignores any new
// array assigned to output.parts.
"command.execute.before": async (input, output) => {
const cmd = input.command;
if (
cmd !== "plannotator-last" &&
cmd !== "plannotator-annotate" &&
cmd !== "plannotator-review"
) return;
output.parts.length = 0;
// input.arguments is the raw tail string from OpenCode's command dispatcher —
// needed so --gate / --json reach the handlers' parseAnnotateArgs.
const event = {
properties: { sessionID: input.sessionID, arguments: input.arguments },
};
if (shouldUseEmbeddedRuntime(workflowOptions.runtime)) {
try {
const embedded = await importEmbeddedRuntime();
const deps = {
client: ctx.client,
htmlContent: getPlanHtml(),
reviewHtmlContent: getReviewHtml(),
getSharingEnabled,
getShareBaseUrl,
getPasteApiUrl,
directory: ctx.directory,
};
const result = await embedded.handleEmbeddedCommand(cmd, event, deps);
if (cmd === "plannotator-last" && result.feedback) {
await embedded.deliverEmbeddedAnnotateMessagePrompt({
client: ctx.client,
sessionId: input.sessionID,
approved: Boolean(result.approved),
feedback: result.feedback,
});
}
return;
} catch (error) {
if (!shouldFallbackAfterEmbeddedError(workflowOptions.runtime, error)) {
throw error;
}
try {
void ctx.client.app.log({
level: "error",
message: `[Plannotator] Embedded runtime unavailable; falling back to CLI: ${error instanceof Error ? error.message : String(error)}`,
});
} catch {}
}
}
if (workflowOptions.runtime === "embedded" && !hasEmbeddedRuntime()) {
try {
void ctx.client.app.log({
level: "error",
message: `[Plannotator] ${getEmbeddedRuntimeError()}`,
});
} catch {}
return;
}
await handleCliCommand({
command: cmd,
client: ctx.client,
sessionId: input.sessionID,
rawArgs: input.arguments ?? "",
cwd: ctx.directory,
bridge: await getBridgeContext(),
});
},
};
if (shouldRegisterSubmitPlan(workflowOptions)) {
plugin.tool = {
submit_plan: tool({
description:
"Submit a plan for user review via line-range edits. First call: pass a single edit with start=1 and your full plan as content (omit end). Subsequent calls after denial: pass targeted edits using the line numbers from the previous response. The tool manages a backing file; you never touch the file directly.",
args: {
edits: tool.schema
.array(
tool.schema.object({
start: tool.schema.number().describe("1-indexed start line (inclusive)"),
end: tool.schema.number().optional().describe("1-indexed end line (inclusive). Omit to replace from start through end of file."),
content: tool.schema.string().describe("Replacement content. Empty string deletes the line range."),
}),
)
.describe("Array of line-range edits to apply to the plan."),
},
async execute(args, context) {
return await executeSubmitPlan({
edits: args.edits,
invokingAgent: context.agent,
sessionId: context.sessionID,
abortSignal: context.abort,
directory: ctx.directory,
workflowOptions,
}, {
reviewPlan: async ({ planContent }) => await runPlanReview({
client: ctx.client,
runtime: workflowOptions.runtime,
planContent,
sharingEnabled: await getSharingEnabled(),
shareBaseUrl: getShareBaseUrl(),
pasteApiUrl: getPasteApiUrl(),
htmlContent: getPlanHtml(),
timeoutSeconds: getPlanTimeoutSeconds(),
abortSignal: context.abort,
cwd: ctx.directory,
bridge: await getBridgeContext(),
}),
resolveTargetAgent: async ({ requestedAgent, directory, delivery }) =>
await resolveValidatedTargetAgent({
client: ctx.client,
targetAgent: requestedAgent,
directory,
delivery,
}),
sendApprovalHandoff: async ({ sessionId, targetAgent, text }) => {
await ctx.client.session.prompt({
path: { id: sessionId },
body: {
agent: targetAgent,
noReply: true,
parts: [{ type: "text", text }],
},
});
},
});
},
}),
};
}
return plugin;
};
export default PlannotatorPlugin;