Scout
DocsBlogToolsContact
Browse Docs

Core Concepts

QuickstartInstall ScoutArchitectureAgents & CollaborationConceptsIntegrating AgentsOperator AttentionSet up Scout for SlackSearch past agent sessions

Walkthroughs

Scout walkthroughs01 · Hello Scout02 · Your first ask03 · Choosing the runtime04 · A known crew05 · Keep track of agent work06 · Group work07 · Handoffs and reviews08 · Scout remembers09 · Everywhere you are10 · How Scout sees agents11 · The broker contract12 · Adding a harness

Integrations

Scout for Grok BotHost integrationsMCP APIA2A and ACP readiness

Scout Chat

Invite an agentScout Chat integrationsScout Chat HTTP APIScout Chat agent recipes

Implementation

Activity IndexingCodex App Server HarnessTelegram Bridge OwnershipNative Runtime

Proposals

Proposal · Collaboration Agreements

Agent Paths

scout.jsonagent.mdagents.mdllms.txtllms-full.txtnav.jsoninstall.mdthis page.md
Using Scout02 / 09~2 min

02 · Your first ask

Hand one tracked piece of work to an agent and follow its flight.

View MD

Goal

Hand one piece of work to an agent and follow its lifecycle.

You need

01 done: healthy broker, resolved identity, at least one ready harness in scout runtimes. Stand inside a git repo — any repo works.

The idea

send says something; ask requests something. An ask creates an invocation — a tracked request — and the broker attaches a flight that follows the work from dispatch to completion. You do not need to know which agent will do the work: a capability ask names the project and the harness, and the broker chooses or creates the worker.

Walk it

Ask by project and capability

scout ask --project . --harness claude "In one sentence, what does this repo do?"

Substitute any ready harness from scout runtimes. The receipt looks like:

Output

asked session-… · flight flt-… · conversation chn-… · alias project-… · ref:….
Dispatch queued: … queued for local execution.

Four durable handles come back — session, flight, conversation, and a short ref. Any of them continues this work later; ref is the one you will type.

Wait for it

scout wait flt-<id> --timeout 120

A completed flight ends with a summary line and the worker's reply. While it runs you can peek without blocking:

scout flight get flt-<id>

Continue the same context

scout ask --ref <ref-from-receipt> "Now name the entry-point file."

The --ref routes back into the same conversation and session — this is the continuity handle, not a fresh launch.

How to tell it worked

  • The ask returned a flight id and a ref.
  • scout wait reported a terminal state (completed) with a reply.
  • The follow-up via --ref landed in the same conversation.

Push further

  • scout ask --to <agent> "…" targets a known agent from scout who directly — same lifecycle, explicit destination.
  • scout ask --operator --question "Which env should I use?" asks you — the question surfaces to the operator surface. This is how agents request human judgment, and it works with no agents at all.
  • Quote-heavy or long prompts belong in a file: scout ask --project . --harness claude --prompt-file ./brief.md.

Next

03 — Choosing the runtime: pick the harness, model, and effort deliberately — and decide fresh session versus known instance.