Talk map · OpenClaw

Context Engineering with Multi-Agent Routing

The practical loop behind Project Transformers: choose the right context, route to the right agent or tool, preserve the outcome, and let the system get smarter without turning the prompt into a junk drawer.

The routing loop

1

Signal arrives

Telegram message, Fizzy card, cron wake, GitHub issue, PR check, browser page, or human nudge.

2

Select context

Load only the relevant memories, conventions, project files, and recent state needed for this decision.

3

Route intent

Pick the execution lane: Optimus, specialist agent, coding harness, first-class tool, cron, or human review.

4

Execute safely

Use scoped tools, repo rules, reversible changes, protected pages, and explicit boundaries for external writes.

5

Persist result

Commit code, update Fizzy, append memory, create PRs, add knowledge, or mark a blocker.

6

Feedback loop

Heartbeats, dispatcher state, status dashboards, and future retrieval turn the output into better future context.

What gets routed where?

SignalContext boundaryRouteDurable output
User askCurrent chat + relevant memory/docs onlyOptimus or specialist agentAnswer, file change, PR, or Fizzy card
Fizzy cardCard description, steps, comments, repo conventionsDispatcher → coding agent / Optimus / humanBranch, PR, card update, review note
GitHub issueIssue labels, repo rules, CI, linked PRsWheeljack / Amp / Codex / OpenCodeImplementation PR with verification
Recurring checkCron payload + current runtime statusIsolated OpenClaw jobSummary, alert, state file, or silence
Knowledge gapMemory + QMD + source filesSearch, retrieve, summarize, compoundKnowledge note or updated convention

Fizzy demo path

1. Card becomes actionable
Human creates/updates a task with steps and assignee.
→
2. Dispatcher evaluates
Fuel, blockers, worktree, routing lane, runtime caps.
→
3. Agent returns artifact
PR, comment, done step, or “needs judgment”.

For the Card #273 demo, start with System Architecture → Communication Architecture to show Fizzy cards, fizzy-pop events, direct messaging, and Telegram topics together, then jump to Ops Dashboard → Fizzy Dispatcher for the live query/fuel/route flow.

Speaker notes

Three punchlines

  • Context engineering is routing. The system decides what context enters the turn before it decides what model answers.
  • Agents are boundaries, not personalities. Each agent owns a domain, memory surface, and permission envelope.
  • Memory is an output channel. If a result matters later, it must become a card, commit, note, or convention.

Where this sits in the existing Viz tour

Start

System Architecture for the full Autobot topology and communication map.

Deepen

Shared Conventions and Memory Management for context selection.

Prove

Operations Dashboard for live cron, fuel, dispatcher, and health posture.