Engineering agent harnesses
18 technical articles on a custom AI concierge and operations platform, with original architecture diagrams and companion code.
Read or print the complete series · Diagram gallery
From an AI concierge to an operations system
Turning guest conversations and hospitality workflows into explicit requirements, authority boundaries and observable outcomes.
Designing the harness: architecture, Hermes and OpenClaw
The runtime boundaries, the selected open-source patterns and the custom control plane built around hospitality operations.
The run contract: orchestration, lifecycle and idempotency
Following one message from authenticated ingress to a durable result, with duplicate delivery, cancellation and retries handled explicitly.
Skills and on-the-fly composition
How two assistants assemble the right instructions, model and tool surface for each turn without becoming a swarm of agents.
Prompt assembly is an engineering pipeline
Stable personas, fresh domain evidence, parallel hydration, prompt ownership and cache-aware composition.
Memory: what was said, what was shown and what is true
Separating conversation history, property knowledge and durable booking or issue state, with scope and freshness preserved.
The model transport and the bounded agent loop
Provider adapters, streamed tool arguments, cancellation, repair passes and the distinction between generated and deliverable text.
Tools as data, execution as a controlled boundary
Compiling tool definitions into handlers, validating backend contracts and keeping model arguments inside authenticated scope.
Authority, approvals and the trust boundary
Separating conversational confirmation from authorization, scoping actions and designing durable human approval.
Booking: let the model converse and the system decide
Search, calendar verification, stable options, fresh checkout links and evidence-backed claims across multiple turns.
From a complaint to work that someone owns
Issue intake, deduplication, staff availability, dispatch ladders, escalation and honest status updates.
One event spine, several ways to deliver the result
Typed runtime events, SSE and synchronous projections, UI components, audit delivery and the boundary between generation and notification.
The control plane: changing behaviour without losing control
Versioned prompts, tools, skills, agents and models; ownership, hot reload, deployability and reproducible releases.
Evaluation starts with the operational scenario
A practical evaluation system for multi-turn booking, issue handling, tool correctness and release decisions.
Judges, human review and regression evidence
Groundedness versus helpfulness, evidence packets, paired comparisons, position bias and human calibration.
The operator console is part of the agent product
An inspectable path from a guest complaint to routing, prompt assembly, tool evidence, cost and a new regression case.
Latency, concurrency and the cost of a completed task
Critical-path analysis, nested limits, cache-aware accounting and load tests that reflect real agent behaviour.
Reliability, release discipline and operational value
Crash boundaries, outbox delivery, configuration consistency, deployment checks and the next steps toward a reusable harness.