From Zero To Hero

A working engineer's briefing · One terminal, one loop, one job that just changed.

• 01The loop flipped — AI writes, you review
• 0230+ hours unattended on real code
• 03The skill is context, not prompts

Knowledge-work briefing · From "where did I save that" to "Claude already filed it."

• 11The unit changed — chat turns out, deliverables in
• 12The desktop is the surface, not the browser
• 13The skill is delegation — picking the right boundary

Patterns, not features — copy, paste, ship. Companion deck to the Code briefing.

• 21I · The terminal IS the IDE (patterns #1–5)
• 22II · Tests are the new prompt (#6–8)
• 23III · Context is a craft (#9–12)
• 24IV · Leverage compounds (#13–20)

Seven layers · One model · Zero trust — defend the model the way you defend the kernel.

• 31The attack surface is the agent's reach
• 32Guardrails are necessary — never sufficient
• 33Identity is the only durable boundary

Sixteen moves you can lift into a runbook before lunch. Companion to the Security briefing.

• 41I · Input boundary (4 patterns)
• 42II · Retrieval & data (4 patterns)
• 43III · Agent & tools (5 patterns)
• 44IV · Runtime & identity (3 patterns)

Cache · Route · Delegate · Plan — the $100 plan doesn't have a limit, your architecture does.

• 51Caching is the lever — all else amplifies it
• 52Opus is a tool, not a default
• 53Plan once, execute cheap

One API, 18 providers, 110+ models — with the IAM you already trust.

• 61Bedrock isn't a model — it's a control plane
• 62Six pillars. Five economic levers. Pick deliberately
• 63Identity & VPC are the moat, not the model

Not for lawyers. For engineers who will be asked "are we compliant?" and need to answer without lying.

• 71Aug 2, 2026 — EU AI Act is fully applicable. Not a drill
• 72Four frameworks. Different jobs. You need all four in some form
• 73Compliance is documentation you can't fake retroactively

Anthropic's canon, framework trade-offs, and the four anti-patterns that kill multi-agent projects.

• 81Most "multi-agent" systems should be a workflow, not agents
• 82The framework is a rounding error — the pattern is the design
• 83Production = HITL gates + observability + budgets

Less product tour, more value. Where it earns its credits, where it burns them, and when to reach for Claude instead.

• 91Manus doesn't answer — it goes and does
• 92The value is in delegation, not conversation
• 93Credits are the constraint — prompt like you're paying

One ladder. Six topics. Three cloud tracks — pick the rung you're on, not the one you wish you were.

• 101The bottleneck isn't access to models — it's a missing rung
• 102Each rung has a build artifact, not a course completion
• 103Cloud track is the last call, not the first

CLAUDE.md · Hooks · Skills · Subagents · Checkpoints · MCP — memory is engineered, not hoped for.

• 111Memory is a hierarchy of files you curate, not chat you hope survives
• 112Four layers: Enterprise · Project · User · Session
• 113Hooks + checkpoints + MCP servers = the difference between L2 and L4

One codebase, three models, three prices — routing beats picking. The right model matches the task's shape.

• 121Sonnet 5 $2/$10 · Opus 4.8 $5/$25 · Fable 5 $10/$50 per 1M tokens
• 122Start at Sonnet 5 — 82.1% SWE-Bench for 40% of Opus's price
• 123~70% Sonnet · ~20% Opus · ~8% Fable 5 · ~2% Haiku — 4× lower bill

Anthropic's Figma-shaped agent · Chat on the left, canvas on the right — from turning it on to shipping to Claude Code.

• 131Design system first — skip it and every project is functional but generic
• 132Live code beats screenshots. Everything beats prompts
• 133/design-sync closes the loop with Claude Code — same account, same usage pool

Your vault becomes the memory, Claude Code becomes the agent — everything captured, cross-linked, ready to be acted on.

• 141The rule of the room: Agents read. Humans write — Claude never fabricates your notes
• 142Four replaceable layers: Markdown files → Git repo → Claude Code → Obsidian UI
• 143Obsidian 1.12 CLI: 54× faster than grep, 60%+ token reduction across skills