services/agents/src/reply.ts · flagon-io/g1t

.prepare(`${selectAgents("a.workspace_id = ?3 AND a.archived_at IS NULL AND a.id <> ?4 AND a.scope = 'workspace'")} ORDER BY a.builtin DESC, a.handle LIMIT 50`)

87

.bind(...periods(now), workspaceId, selfId)

88

.all<Row>();

89

return rows.results.map((row) => {

90

const agent = toAgent(row, now);

91

return {

92

handle: agent.handle,

93

display_name: agent.display_name,

role: agent.role,

title: agent.title,

team: agent.team,

department: agent.department,

98

responsibilities: agent.responsibilities,

99

status: agent.status,

100

spent_month_micros: agent.spent_month_micros,

101

monthly_micros: agent.budget.monthly_micros,

};

});

}

/** The agent's latest sessions in this conversation, one line each. */

107

async function sessionsHere(db: D1Database, agentId: string, channelId: string): Promise<string | null> {

108

const rows = await db

109

.prepare(

110

"SELECT id, title, status, summary, created_at FROM agent_sessions WHERE agent_id = ? AND channel_id = ? AND parent_id IS NULL ORDER BY created_at DESC LIMIT 5",

111

)

112

.bind(agentId, channelId)

113

.all<{ id: string; title: string; status: string; summary: string | null; created_at: string }>();

114

if (!rows.results.length) return null;

115

return rows.results

116

.map((s) => `- "${s.title}" (${s.status}, ${s.created_at.slice(0, 10)})${s.summary ? `: ${s.summary.replace(/\s+/g, " ").slice(0, 300)}` : ""}`)

.join("\n");

}

/**

* Consulting a colleague:

122

* the colleague answers in a nested turn that posts nothing, with the same

123

* audience (so it can read no more than the conversation may), the same

124

* asker, one hop further, its own routing limits, on the same model

125

* session, billed to this reply. A compact card in the thread says who

126

* asked whom. Consults don't nest: the colleague can't consult in turn.

127

*/

128

function consulting(input: {

db: D1Database;

row: Row;

delivery: DeskWork;

send: Send;

policy: Policy;

ownModel: boolean;

sessionModel: string | null;

136

own: string | null;

137

ports: (consult: ToolPorts["consult"]) => ToolPorts;

hops: number;

now: Date;

surface: Surface;

asker: { name: string; display_name: string | null; access: NonNullable<AgentDelivery["asker"]> | null };

142

spent: { tokens: Tokens; cost: number };

143

}): { ask: ToolPorts["consult"]; attach(box: ToolBox): void } {

144

let parent: ToolBox | null = null;

145

const noNesting: ToolPorts["consult"] = async () => ({ ok: false, message: "Consults don't nest: answer with what you have." });

146

const ask: ToolPorts["consult"] = async (handle, question) => {

147

const { row, delivery } = input;

148

const colleague = await input.db

149

.prepare("SELECT * FROM agents WHERE workspace_id = ? AND handle = ? AND archived_at IS NULL AND scope = 'workspace'")

150

.bind(row.workspace_id, handle)

151

.first<Row>();

152

if (!colleague || colleague.id === row.id) return { ok: false, message: `There is no other agent called @${handle} here.` };

153

// No ping-pong: never back to the agent that sent this work.

154

if ((delivery.chain ?? []).at(-1) === colleague.id) return { ok: false, message: `@${handle} sent you this work; answer with what you have.` };

155

if (input.hops + 1 > CHAT_MAX_HOPS) return { ok: false, message: "This request has been passed along too many times; answer with what you have." };

156

const definition = definitionOf(colleague);

157

const allowed = allowedProviders(definition.routing, input.own);

158

if (input.ownModel ? !allowed.own : !allowed.hosted) return { ok: false, message: `@${handle} can't use the model this conversation runs on.` };

159

const model = replyModel(

160

input.policy,

161

{ ...definition.routing, pinned: input.ownModel ? definition.routing.pinned : null },

162

{ named: input.ownModel ? input.sessionModel : null },

163

);

164

const tools = parent

165

? parent.forColleague(input.ports(noNesting), { agentId: colleague.id, notConsult: [colleague.handle, row.handle], hops: input.hops + 1, maxHops: input.hops + 1 })

166

: null;

167

const system = systemPrompt({

168

agent: { ...definition, id: colleague.id },

169

workspace: delivery.workspace,

170

channel: { kind: delivery.channel_kind, name: delivery.channel_name },

171

asker: input.asker,

172

today: input.now,

173

tools: tools ? { code: tools.definitions().some((tool) => tool.name === "read_file") } : null,

174

consultedBy: row.handle,

175

});

176

const result = await runTurn(input.send, {

177

model: model.model,

178

system,

179

messages: [{ role: "user", content: `@${row.handle} (agent) asks you: ${question}` }],

180

tools,

181

price: input.ownModel ? null : model.price,

182

});

183

input.spent.tokens = addTokens(input.spent.tokens, result.tokens);

184

input.spent.cost += input.ownModel ? 0 : result.cost;

185

const answer = result.text || "(no answer)";

186

const exchange = `Q: ${question}\nA: ${answer}`;

187

// The exchange, collapsed, in the thread. No @ in its text: it wakes nobody.

188

await input.surface

189

.post(`${row.display_name} asked ${colleague.display_name}`, {

190

kind: "consult",

191

title: `${row.display_name} asked ${colleague.display_name}`,

192

detail: exchange.length > 1000 ? `${exchange.slice(0, 1000)}…` : exchange,

state: null,

href: null,

})

.catch((error: unknown) => console.error("agents: a consult card was not posted", String(error)));

197

return { ok: true, colleague: colleague.handle, answer };

};

return {

ask,

attach(box) {

parent = box;

},

};

}

type Outcome = {

status: "replied" | "steered" | "blocked" | "skipped" | "failed";

209

error?: string | null;

210

reply_id?: string | null;

211

model?: string | null;

212

tier?: string | null;

tokens?: Tokens;

cost?: number;

charged?: number;

/** The effort level it ran at. */

217

effort?: string | null;

};

/**

* What a desk is handed: a message to answer, or (`hello`) the agent's

222

* first message to the person who made it, in the DM that just opened.

223

*/

224

export type DeskWork = AgentDelivery & { hello?: boolean };

225

226

/** How a limit's refusal reads in chat, in the agent's voice. */

227

function noticeFor(reason: string, message: string): string {

228

if (reason.startsWith("budget_") || reason === "workspace_agent_budget" || message.startsWith("My settings") || message.startsWith("I ")) return message;

229

if (reason === "no_model") return `I can't reply yet: ${message}`;

230

return `I can't reply right now: ${message}`;

231

}

232

233

/** The session a message in this thread is for: a reply under one of this agent's session cards. */

234

async function steeredSession(db: D1Database, agentId: string, threadRoot: string | null): Promise<SessionRow | null> {

235

if (!threadRoot) return null;

236

const found = await db.prepare("SELECT id FROM agent_sessions WHERE agent_id = ? AND card_message_id = ?").bind(agentId, threadRoot).first<{ id: string }>();

237

return found ? sessionRow(db, found.id) : null;

}

/**

* Answers `delivery` as its agent. Never throws: every way it ends is

242

* recorded on the reply's row. A message handed over twice is answered

243

* once.

244

*/

245

export async function reply(env: ReplyEnv, delivery: DeskWork, now = new Date()): Promise<void> {

246

const db = env.DB;

247

const row = await db.prepare("SELECT * FROM agents WHERE id = ?").bind(delivery.agent_id).first<Row>();

248

if (!row || row.archived_at || row.workspace_id !== delivery.workspace_id) return;

249

const id = newId("arp", now.getTime());

250

const claimed = await db

251

.prepare(

252

`INSERT INTO agent_replies (id, agent_id, workspace_id, channel_id, message_id, asked_by, asked_by_username, channel_name, agent_version, status, created_at)

253

VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, 'working', ?)

254

ON CONFLICT (agent_id, message_id) DO NOTHING RETURNING id`,

255

)

256

.bind(id, row.id, row.workspace_id, delivery.channel_id, delivery.message_id, delivery.asked_by, delivery.asker?.username ?? null, delivery.channel_name, row.version, now.toISOString())

257

.first<{ id: string }>();

258

if (!claimed) return;

259

260

const surface = surfaceFor(env.CHAT, delivery);

261

// 👀 as soon as the desk has it, while everything else goes on.

262

const acknowledging = surface.acknowledge();

263

const slug = delivery.workspace.toLowerCase();

264

265

// What the answer used, once there is one: counted however the reply ends.

266

let usage: Partial<Outcome> = {};

267

// What the tools did, for the audit table, and who the audience was.

268

let toolCalls: ToolCall[] = [];

269

let audienceHash: string | null = null;

270

271

const finish = async (outcome: Outcome) => {

const statements = [

db

.prepare(

`UPDATE agent_replies SET status = ?, error = ?, reply_id = ?, model = ?, tier = ?, input_tokens = ?, output_tokens = ?,

276

cost_micros = ?, charged_micros = ?, tool_count = ?, finished_at = ?, effort = ? WHERE id = ?`,

)

.bind(

outcome.status,

outcome.error?.slice(0, 1000) ?? null,

281

outcome.reply_id ?? null,

282

outcome.model ?? null,

283

outcome.tier ?? null,

284

(outcome.tokens?.input ?? 0) + (outcome.tokens?.cacheRead ?? 0) + (outcome.tokens?.cacheWrite ?? 0),

285

outcome.tokens?.output ?? 0,

286

outcome.cost ?? 0,

287

outcome.charged ?? 0,

288

toolCalls.length,

289

new Date().toISOString(),

290

outcome.effort ?? null,

291

id,

292

),

293

// Every tool call: what was asked, for whom, and whether it was read or withheld.

294

...toolCalls.map((call, n) =>

295

db

296

.prepare(

297

`INSERT INTO agent_tool_calls (id, reply_id, agent_id, workspace_id, asked_by, tool, args, audience_hash, outcome, bytes, created_at)

298

VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)`,

299

)

300

.bind(`${id}_${n}`, id, row.id, row.workspace_id, delivery.asked_by, call.tool, call.args, audienceHash ?? "", call.outcome, call.bytes, now.toISOString()),

301

),

302

];

303

await db.batch(statements);

304

// ✅ when it answered; 👀 taken back after a notice, an apology or nothing.

305

await acknowledging;

306

await surface.settle(outcome.status === "replied" || outcome.status === "steered" ? "done" : "withdrawn");

307

};

308

309

/** Says once, in this conversation, why the agent cannot answer; a repeat within hours is kept back. */

310

const notice = async (message: string, reason: string) => {

311

// An agent's first words are never an excuse: without a model, a fixed hello.

312

if (delivery.hello) {

313

const posted = await surface.post(fixedHello(row, delivery.asker?.username ?? null)).catch(() => null);

314

return await finish({ status: "blocked", error: reason, reply_id: posted });

315

}

316

const since = new Date(now.getTime() - NOTICE_QUIET_MS).toISOString();

317

const recent = await db

318

.prepare(

319

`SELECT 1 FROM agent_replies WHERE agent_id = ? AND channel_id = ? AND status = 'blocked' AND error = ?

320

AND reply_id IS NOT NULL AND created_at > ? AND id <> ? LIMIT 1`,

321

)

322

.bind(row.id, delivery.channel_id, reason, since, id)

323

.first();

324

const posted = recent ? null : await surface.post(message).catch(() => null);

325

await finish({ status: "blocked", error: reason, reply_id: posted });

};

try {

// A reply under one of its session cards steers that session; nothing is answered here.

330

const steered = delivery.hello ? null : await steeredSession(db, row.id, delivery.thread_root);

331

if (steered) {

332

const history = await surface.history(HISTORY_LIMIT).catch(() => [] as SurfaceMessage[]);

333

const message = history.find((m) => m.id === delivery.message_id);

334

if (message?.body.trim()) {

335

await steer(env as unknown as SessionEnv, steered, message.author.name, message.body);

336

return await finish({ status: "steered" });

}

}

// Read the conversation while showing that the agent is on it.

341

// A hello has no conversation yet: it is asked to introduce itself.

342

// And who is here: said every turn, so the agent knows who reads it and who doesn't.

343

// And its teams, from their pages: who leads, who owns what, who is around.

344

const [, history, conversationHere, teamsHere] = await Promise.all([

345

surface.typing(),

346

delivery.hello ? Promise.resolve([]) : surface.history(HISTORY_LIMIT),

347

delivery.hello ? Promise.resolve(null) : surface.conversation(),

348

loadTeams(env, slug, row.workspace_id, { id: row.id, team: row.team ?? null }).catch(() => null),

349

]);

350

// A personal agent answers only its member, in the DM of the two of them (access.ts).

351

const refused = personalRefusal(row, {

352

asked_by: delivery.asked_by,

353

channel_kind: delivery.channel_kind,

354

members: conversationHere ? conversationHere.people + conversationHere.agents : null,

355

});

356

if (refused) return await notice(refused, "personal_agent");

357

const conversation = delivery.hello ? [{ role: "user" as const, content: helloAsk(delivery.asker?.username ?? null) }] : turns(history, row.id);

358

if (!conversation.length) return await finish({ status: "skipped", error: "nothing to answer" });

359

const author = askerIn(history, delivery);

360

const askerName = delivery.asker?.username ?? author?.name ?? null;

361

// Every agent knows its colleagues; @g1t also steps up a tier to decide

362

// who gets the work in a long thread.

363

const specialists = await team(db, row.workspace_id, row.id, now);

364

const definition = definitionOf(row);

365

// Its effort setting decides where the reply starts and how hard the model reasons.

366

const plan = effortPlan(effortOf(definition.routing), row.builtin ? orchestratorTier(history.length, specialists.filter((a) => a.handle !== "g1t").length) : REPLY_TIER);

367

const hops = Math.max(0, Math.floor(delivery.hops || 0));

368

const chain = delivery.chain ?? [];

369

const sender = chain.length ? await db.prepare("SELECT handle FROM agents WHERE id = ?").bind(chain[chain.length - 1]).first<{ handle: string }>() : null;

370

const asker = { name: askerName ?? "someone", display_name: author?.display_name ?? null, access: delivery.asker ?? null };

371

let posted: string | null = null;

372

let spinOffs = 0;

373

374

const done = await metered(env, { row, payer: row, slug, task: "reply", start: plan.start, effort: plan.level, askerName, person: delivery.asker?.username ?? null, teams: teamsHere }, async (model) => {

375

// What colleagues consulted along the way used: billed to this reply.

376

const consulted = { tokens: NO_TOKENS, cost: 0 };

377

let toolbox: ToolBox | null = null;

378

let place: RecallPlace = { channel_id: delivery.channel_id, kind: delivery.channel_kind === "dm" ? "dm" : "private", people: [delivery.asked_by] };

379

if (!delivery.hello) {

380

try {

381

const audience = await Audience.build(slug, delivery.asked_by, audiencePorts(env, slug, delivery.channel_id));

382

place = { channel_id: delivery.channel_id, kind: audience.kind, people: audience.shared ? [delivery.asked_by] : audience.members.map((m) => m.id) };

383

const ports = (consult: ToolPorts["consult"]) => toolPorts(env, slug, row.workspace_id, delivery.channel_id, consult, row.id);

384

audienceHash = audience.hash;

385

const consult = consulting({

db,

row,

delivery,

send: model.send,

policy: model.policy,

391

ownModel: model.ownModel,

392

sessionModel: model.sessionModel,

own: model.own,

ports,

hops,

now,

surface,

asker,

spent: consulted,

});

const where = delivery.channel_kind === "dm" ? "a direct message" : `#${delivery.channel_name ?? "a channel"}`;

402

const actions = actionPorts(env as unknown as SessionEnv, {

403

agent: row,

404

place,

405

source: { kind: "message", ref: delivery.message_id, label: askerName ? `@${askerName} in ${where}` : where, channel_id: delivery.channel_id },

406

asker: { id: delivery.asked_by, username: askerName },

407

workspace: slug,

408

postCard: (card) => surface.post("", card).catch(() => null),

409

// Real work becomes a session, with its card in this conversation.

410

spinOff: async (title, goal) => {

411

if (spinOffs >= MAX_SPIN_OFFS) return { ok: false, message: "You've started enough sessions from this message." };

412

spinOffs++;

413

const session = await startSession(env as unknown as SessionEnv, {

agent: row,

kind: "chat",

title,

goal: `${askerName ? `@${askerName}` : "Someone"} asked in ${where}:\n\n${goal}`,

418

workspace: slug,

419

channel_id: delivery.channel_id,

420

channel_kind: delivery.channel_kind,

421

channel_name: delivery.channel_name,

422

thread_root: delivery.thread_root,

423

message_id: delivery.message_id,

424

asked_by: delivery.asked_by,

425

asked_by_username: askerName,

426

asker: delivery.asker ?? null,

chain: [...chain],

hops,

});

const cap = session.cap_micros ? ` with a cap of ${dollars(session.cap_micros)}` : "";

431

return { ok: true, message: `Started the session "${session.title}"${cap}. Its card is in the conversation and it reports back there. Tell them in a sentence; don't do the work here.` };

432

},

433

});

434

// Work handed to a colleague: the only way an agent gets another working (handoff.ts).

435

const viewer = audience.asker;

436

actions.handOff = handOffPort({

437

self: { id: row.id, handle: row.handle },

438

chain,

439

asker: delivery.asker ?? null,

440

agent: async (handle) => {

441

const found = await db

442

.prepare(selectAgents("a.workspace_id = ?3 AND a.handle = ?4 AND a.archived_at IS NULL AND a.scope = 'workspace'"))

443

.bind(...periods(now), row.workspace_id, handle)

444

.first<Row>();

445

if (!found) return null;

446

const agent = toAgent(found, now);

447

return { id: agent.id, handle: agent.handle, display_name: agent.display_name, builtin: !!found.builtin, status: agent.status };

448

},

449

person: async (handle) => {

450

if (!viewer) return false;

451

const members = await identityClient(env.IDENTITY).listMembers(slug, viewer);

452

return members.ok && members.value.some((m) => m.username.toLowerCase() === handle);

453

},

454

handOff: (colleagueId, brief) => surface.handOff(colleagueId, brief),

455

});

456

toolbox = new ToolBox(

457

audience,

458

ports(consult.ask),

459

{ agentId: row.id, notConsult: [row.handle, ...(sender ? [sender.handle] : [])], hops, maxHops: CHAT_MAX_HOPS },

[],

actions,

);

consult.attach(toolbox);

464

} catch (error) {

465

// Without an audience nothing may be read: the reply goes on with this conversation only.

466

console.error("agents: no audience for a reply, so no tools", row.id, String(error));

467

}

468

}

469

// What people said last, for recalling what the workspace's artifacts say about it.

470

const said = [...history].reverse().filter((m) => m.author.kind === "user").slice(0, 3).map((m) => m.body);

471

const [facts, recent, passages] = delivery.hello

472

? [[], null, []]

473

: await Promise.all([

474

recall(db, row.id, place).catch(() => []),

475

sessionsHere(db, row.id, delivery.channel_id).catch(() => null),

476

toolbox ? toolbox.recall(recallQuery(said), definition.reading ?? []) : Promise.resolve([]),

]);

const system = [

systemPrompt({

agent: {

...definition,

id: row.id,

// @g1t's job is fixed; what the workspace wrote is added to it.

484

instructions: row.builtin ? orchestratorInstructions(specialists.filter((a) => a.handle !== "g1t"), definition.instructions) : definition.instructions,

485

},

486

workspace: delivery.workspace,

487

channel: { kind: delivery.channel_kind, name: delivery.channel_name },

488

asker,

489

today: now,

490

tools: toolbox ? { code: toolbox.definitions().some((tool) => tool.name === "read_file") } : null,

491

// @g1t's team is in its job; everyone else is told who their colleagues are.

492

colleagues: row.builtin ? null : rosterLines(specialists),

493

recentSessions: recent,

494

conversation: conversationHere,

495

teams: teamsSection(row.id, teamsHere, now),

496

canHandOff: !!toolbox?.definitions().some((tool) => tool.name === "hand_off"),

497

handedOffBy: sender?.handle ?? null,

498

skills: skillsSection(definition.skills_off, toolbox?.definitions().map((tool) => tool.name) ?? []),

499

}),

500

memorySection(facts),

501

recallSection(passages),

]

.filter(Boolean)

.join("\n\n");

toolCalls = toolbox?.calls ?? [];

506

const answer = await runTurn(model.send, { model: model.model.model, system, messages: conversation as ModelMessage[], tools: toolbox, price: model.ownModel ? null : model.model.price, effort: model.effort });

507

// Post as soon as there is an answer; the bill is settled after.

508

// It wakes nobody, @mentions or not: colleagues get work only through hand_off.

509