Menu
Contact Adam
Home

AgentRecipes

Practical patterns for building real agents: memory, tools, workflow state, review loops, and learning systems.

This section is for builders who already know the model is only one part of the system. The useful questions are usually more structural: what should live in memory, what should live in state, what should be logged, and where human review should sit.

These recipes are meant to be implementation-oriented. They are not prompt tricks. They are patterns that tend to hold up better once an agent moves past the demo stage and starts doing repeated work.

What you'll find here:

  • How to separate agent identity from workflow state
  • How to compress session history without losing continuity
  • Where to add human approval gates in real workflows
  • How to keep tool-enabled agents from collapsing into prompt sprawl
  • What to log if you want your system to improve from real usage

Memory

Memory

Identity vs. State

How to separate durable agent identity from changing workflow state so your agent stays consistent in real operation.

5 minBuilders working on agents that run repeatedly across workflows, approvals, queues, or customer conversations.
Memory

Session Summaries That Actually Help

A practical pattern for compressing conversation history into something an agent can reuse without dragging old noise into every run.

6 minAgents that stay active across days or weeks and need continuity without huge prompts.

Workflow

Workflow

Human Approval Gates

How to add human review at the right moment without turning the agent into a useless draft machine.

5 minOutbound, support, finance, ops, or any agent that can trigger a real-world action.

Tooling

Tooling

Tool Use Without Prompt Sprawl

A cleaner pattern for tool-enabled agents: small core rules, explicit tool surfaces, and task-specific context only when needed.

5 minCoding agents, operator agents, and assistant systems with many tools or integrations.

Learning

Learning

Logging Agent Outcomes

What to log from agent runs so you can improve prompts, retrieval, evaluations, and eventually training without over-instrumenting the system.

6 minTeams that want their agent systems to get measurably better over time.

Need an agent system built around a real workflow?

If you already know the use case and want help turning it into something production-ready, that's the kind of work I do.

Ask directly or call/text: (603) 748-4982

All guides | How it works | Pricing