Repository navigation
Where do I find it
A lookup table across all four surfaces. If you know what you want but not where it is, start here.
| I want… | It is here |
|---|---|
| What an agent actually is, vs a workflow | Module 00 · Autonomy Ladder |
| How LLMs work, without maths | Module 01 · LLM Intuition Bank |
| What transfers off AWS | Portability matrix |
| The agentic PDLC, as one simulation | SkyWays, the agentic manual · ideas thread |
| A definition of a term | Glossary |
| Which AWS service does what | AWS service map |
| I want… | It is here |
|---|---|
| The four phases and what each one owes the next | The Agentic PDLC |
| Which loop closes where, and who owns it | The Eight Loops |
| The five gates, and which ones are mine | Gates & Governance |
| The documents owed at each hand-off | The Evidence Pack |
| My whole role, end to end, with a template per step | PM · SA · Engineering · QA · DevOps |
| A prompt I can paste | Every prompt |
| A template for an artefact | Every template |
| What I am accountable for, and what I must not touch | PM · Architect · Engineering · QA · DevOps · Sponsor |
| How to derive an acceptance bar | Formulas & Calculators · How to Prove the Bar |
| Whether a score has actually proven the bar | How to Prove the Bar |
| Why the token bill left its estimate | How to Control the Token Bill |
| Where a cap should live | How to Hold the Security Boundary |
| How to run the postmortem | How to Run a Missing-Control Postmortem |
| How to run the NFR workshop | How to Run an NFR Workshop |
| How many agents this needs | How to Design an Agent on Paper · Decision Trees |
| How to unblock a review queue | How to Review by Risk Band |
| An intuition I can carry into a case this manual never covered | Mental Models |
| The whole thing explained to a CEO | The operating protocol |
| A term this manual uses oddly | Glossary |
| A case from my own industry | Scenario Library |
| Practice questions, with answers | Exercises & Answers |
| Whether a number here is a standard or a default | Sources & Confidence |
A how-to is a procedure with an owner, a phase and a loop it closes. Each opens with an At a glance block saying when to reach for it and what you leave with. Pick by what is going wrong, not by what sounds relevant.
| Reach for it when | How-to | Owner | Closes |
|---|---|---|---|
| Somebody is about to open an editor and nobody has written down what the thing may do | Design an Agent on Paper | Solution architect | Spec |
| The NFRs are adjectives, and trade-offs get settled by seniority | Run an NFR Workshop | Solution architect | Requirements |
| Somebody has already picked the answer, and the decision must survive them leaving | Choose Build, Buy or Borrow | Solution architect | Decision |
| A score is being offered as proof | Prove the Bar | QA lead | Trust |
| The plan is ordered by priority, and day three needs something from day seven | Cut Sprints into Bolts | Engineering lead | Delivery |
| The queue is days long and the policy is two seniors on everything | Review by Risk Band | Engineering lead | Delivery |
| You cannot name the line of code that refuses | Hold the Security Boundary | Engineering lead, with the architect | — (the control set every loop assumes) |
| The bill left its estimate and traffic did not | Control the Token Bill | Engineering lead | Cost |
| The draft action list contains a string rather than a control | Run a Missing-Control Postmortem | QA lead | Incident |
| I want… | It is here |
|---|---|
| To write the agent loop by hand | Module 05 · AGL-01 |
| Converse API syntax | Cheat sheet |
| Strands / LangGraph / AgentCore syntax | Quick reference |
| To build RAG properly |
Module 10 · ragkit
|
| To deploy an agent as a service | Module 11 · deploy how-to |
| Graded practice | L.A.B. Simulator — or one click in Codespaces |
| The ideal lifecycle for a project like this | Agentic PDLC · Lifecycle Reference — artefacts and gates mapped to AIDD, BMAD, AI-DLC |
| What is active in the repo right now | Repo Pulse |
| Graded practice with nothing installed | Simulator Arena |
| Practice with other people's answers | Discussion index |
| I want… | It is here |
|---|---|
| To decide if this should be an agent | Autonomy Ladder · PDL-01 |
| To choose a multi-agent topology | Handoff Multiplier · how-to |
| To decide build vs buy | Playbook |
| To run a design review | Playbook · Scorecard |
| To write acceptance criteria | How-to |
| A PRD template | Seven sample PRDs |
| I want… | It is here |
|---|---|
| A specific error message | Error Index |
| The AWS-specific errors | Troubleshooting |
| To debug from a symptom | Failure Signature Catalog |
| An incident procedure | Runbooks |
| Things that break silently | Silent Degradation Watchlist |
| I want… | It is here |
|---|---|
| The two-line check for the repeating-tool-call bug | agent_history_invariant.py |
| A tool return contract the model cannot misread | honest_tool_result.py |
| A release gate in 40 lines that can say no | release_gate.py |
| An H× calculator for a design review | handoff_multiplier.py |
| I want… | It is here |
|---|---|
| To prepare for an interview | Interview guides — five roles, both sides |
| To hire for this | Hiring guide |
| To teach this to a team | Training frameworks playbook |
Not here? Add a row — anyone can edit. Or ask in Q&A and the answer becomes a row.
This wiki is the connective layer. The repository is canonical and versioned; these pages are the cross-surface maps and the notes that change too often to live in a reviewed file. Anyone can edit — see Editing this wiki.
▶ Start here · 55 lessons, on the site ↗
- Getting started · 2
- Agentic PDLC fundamentals · 9
- Methods decoded · 7
- Running delivery · 13
- By role · 9
- Teams and organisation · 4
- Practice · 4
- Interviews and careers · 7
The method
The Agentic PDLC
The Eight Loops
Gates & Governance
The Evidence Pack
The journey · end to end
Product manager
Solution architect
Engineering lead
QA lead
DevOps & platform
The method · by role
Product manager
Solution architect
Engineering lead
QA lead
DevOps and platform
Sponsor
How to
Run an NFR workshop
Design an agent on paper
Build, buy or borrow
Cut sprints into bolts
Review by risk band
Prove the bar
Control the token bill
Hold the security boundary
Run a missing-control postmortem
Reference
Mental Models
Glossary
Formulas & Calculators
Decision Trees
Scenario Library
Exercises & Answers
Anti-Patterns
Sources & Confidence
Where do I find…?
The 16 modules
- Agentic Foundations
- LLM Intuition and the AWS Bridge
- Amazon Bedrock Essentials
- Amazon Bedrock Agents
- Agent Builder, Knowledge Bases and Guardrails
- The Agent Loop: No Framework to Strands
- Strands Foundations: Tools, Memory and MCP
- Multi-Agent Patterns with Strands
- LangChain and LangGraph
- LLM Memory Mechanics
- RAG, OpenSearch and LiteLLM
- Amazon Bedrock AgentCore
- A2A and A2UI: Agent Interoperability
- Agentic QA and Evaluation
- End-to-End Production Pipeline
- Agentic Product Lifecycle
Learning paths · Field guide
Study Plans
Eight sessions
Living notes
Error Index
Community Answers
Field Notes
Model & Region Notes
Cost Log
Scoreboard 🧪
The project
Roadmap
Editing this wiki
Maintainer Runbook
📘 The manual · live
🛫 Ninety Days, the simulator
🧰 The workbench
👔 For leadership
🧠 Mental models
Templates · Prompts
Frameworks & acronyms
The repository · Setup
Pinned discussions
Discussion map — where to post what
Exercise & lab index
Simulator Arena — get a lab graded
Pitch an idea
All discussions →
