Build an agent that remembers, acts for you, and can be trusted.
Eight lessons that build one standing agent end to end: a memory layer it reads and writes, retrieval that stays clean, tool integrations that never see your credentials, a skill library, a scheduled loop, guardrails, and a deployment that runs without you.
You finish holding something.
You leave with a running personal agent — not a demo. It remembers across sessions, reaches your tools through a wrapper that isolates secrets, runs on a schedule, and is small enough that you can audit every line.
The artifact
A standing agent you own: memory files, a search index, a skill library, guardrail hooks, and a heartbeat, deployed and running.
The loop you learn
Remember → Retrieve → Act → Report
Every lesson has the same shape.
You know what you are buying before you buy it.
01The problem
What breaks in a standing agent without this piece, shown concretely.
02The build
The files, in full, with the reasoning for each design choice.
03The check
How you prove the piece works before moving on.
04The deliverable
The named artefact now sitting in your repository.
Why it is built this way.
Three decisions that shaped every lesson.
Every lesson ships a deliverable
No lesson ends in theory. Each one leaves a working piece on disk, and the next lesson builds on it directly.
Small enough to audit
The whole system stays a few hundred lines you wrote. You can read all of it, which is the only reason you can trust any of it.
Least privilege by default
The agent gets read access first and write access only where you deliberately grant it. Credentials never enter the model's context.
3 modules, 10 entries.
The whole path is public. Nothing is hidden until you pay — only the bodies are gated.
MOD 01Memory
The layer everything else depends on: what the agent knows about you, and how it stays useful as it grows.
- 01What a Standing Agent IsRead · 27m · concise path · 4 quiz · free · updated 2026-09-15
You end with: An architecture sketch and a requirements file describing the agent you specifically want.
- 02The Memory LayerRead · 22m · concise path · 4 quiz · updated 2026-09-13
You end with: Working memory files plus the read/write protocol the agent follows every session.
- 03Retrieval and CurationRead · 20m · concise path · 4 quiz · updated 2026-09-27
You end with: Local hybrid search over your memory, plus a reflection pass that prunes what went stale.
MOD 02Reach
Give the agent hands — tools and skills — without giving it your credentials.
- 04Tools and IntegrationsRead · 19m · concise path · 4 quiz · updated 2026-09-13
You end with: One real integration behind a wrapper that keeps API keys out of the model's context.
- 05The Skill LibraryRead · 28m · concise path · 4 quiz · updated 2026-09-15
You end with: Three skills in a shared format, loaded by progressive disclosure.
- 06Build a Minimal Standing AgentProject · 12m · updated 2026-09-16
You end with: A minimal standing agent that remembers, retrieves, and acts through one tool.
MOD 03Autonomy and trust
Let it run on its own — and make that a safe thing to have done.
- 07The Scheduled LoopRead · 19m · concise path · 4 quiz · updated 2026-09-13
You end with: A heartbeat that reports when something matters and stays silent otherwise.
- 08Guardrails and HardeningRead · 20m · concise path · 4 quiz · updated 2026-09-13
You end with: Hook-based guardrails, a command allow-list, and a prompt-injection test that fails loudly.
- 09Deployment and EvolutionRead · 22m · concise path · 4 quiz · updated 2026-09-13
You end with: The agent running on a schedule outside your laptop, with an evolution log that closes the loop.
- 10Audit Your Agent's Blast RadiusProject · 10m · updated 2026-09-16
You end with: A written audit of your agent's blast radius, with the gaps closed.
Start with the free lesson.
Engineers who have used a coding agent and now want one that works on their behalf between sessions.
- One-time price, lifetime updates
- Every course included — this one and the next
- 30-day money-back guarantee
- No card required to start