Just took Anthropic's Architect Foundations exam
A user recently completed Anthropic's Architect Foundations exam and shared insights for others preparing for it. They noted the scarcity of preparation materials and provided a link to their Anki deck for spaced repetition. The user found the exam beneficial for those regularly building agentic systems with Claude, as it encourages a deeper understanding of underlying mechanisms beyond mere prompt engineering.
Time & source
Times shown in UTC
Display time zone: UTC
Local time zone unavailable; showing UTC.
PublishedOffset at this time: UTC+0Sep 12, 2026, 15:54 UTC
IngestedOffset at this time: UTC+0Sep 12, 2026, 18:01 UTC
- Published
- Sep 12, 2026, 15:54
- Ingested
- Sep 12, 2026, 18:01
- Source type
- Dev community
- Tier
- Community
- Source status
- Healthy
Tier is a per-source editorial setting, not a per-item score.
Register to choose your default translation language in settings.
So I just finished the Architect Foundations exam and figured I'd drop some notes here since I couldn't find much when I was prepping.
TL;DR: don't go in thinking this is a "read the docs for an hour" cert. It's dense as hell, tons of reading, and it actually tests whether you understand how agentic systems work, not just whether you've used Claude Code a few times.
Stuff that showed up (from memory, no guarantees lol):
Agents/tools
- structured outputs (nullable vs optional, formatting rules)
- writing tool descriptions + few-shot examples that actually help tool selection
- chaining tool calls (using the ID/output from one call as input for the next)
- when to retry a failed tool call vs just surfacing the error
- MCP structured messages / structured error responses
- where MCP context actually gets loaded in a session
Orchestration
- subagents, context flooding, resuming sessions
- forking a session
- plan → execute → review vs direct execution vs "interview" style, and which tasks actually benefit from planning first
- the tool loop itself (tool_use, tool_result, end_turn, how a turn resolves)
Config/control
- CLAUDE.md vs PreToolUse hooks vs settings.json permissions - when to use which
- managing context window over long sessions
- prompt-based guardrails vs deterministic/hardcoded safety
- grep vs glob vs read
- resuming a session with a proper "here's what changed" summary
Basically expect a LOT of text as your "source material," it's not just watching a few videos and calling it a day.
Stuff that helped me study:
- My Anki deck (spaced repetition): https://ankiweb.net/shared/info/1419215955