Skip to content

resource

Runnable Python AI Agent Chapter Checkpoints

Run the supplied reference checkpoints for the drafted chapters. Follow the locked environment setup and distinguish offline mechanism checks from optional live observations.

Updated: 2026-09-09 · Status: DRAFT

Run from the complete repository root using the frozen environment. These demonstrations accompany the manuscript. The ownership guide distinguishes learner-built definitions from supplied runtime behavior. New Chapters 4, 7 and 13 have construction briefs, not completed checkpoints.

ChapterManuscriptCheckpoint
1Make the first model call for Lucych01.py
2Give the agent reliable shop toolsch02.py
3Build the model and tool loopch03.py
4Build durable state with SQLitePLANNED — no executable checkpoint
5Remember across conversationsch05.py
6Reuse a tested opening procedurech06.py
7Build a durable work inbox and report outboxPLANNED — no executable checkpoint
8Talk to the agent from your phonech08.py
9Wake up for schedules and stock eventsch09.py
10Ask permission before spendingch10.py
11Survive the ambiguous supplier orderch11.py
12Recover work after a process crashch12.py
13Connect an external tool with MCPPLANNED — no executable checkpoint
14Isolate tools and untrusted contentch14.py
15Measure whether the agent helpsch15.py
16Improve behavior with evaluated changesch16.py
17Delegate one bounded taskch17.py
18Deploy and maintain the agentch18.py
19Lucy leaves the shop for a daych19.py

The final accelerated day retains two databases, a readable report and JSON evidence:

bash
uv run --python 3.14 python book/textbook/checkpoints/profrod_sovereign_agent_ch19_integrated_shop_day_checkpoint.py --output /tmp/lucy-day-first-run

Use a fresh output path for each retained run; the checkpoint refuses to overwrite it. The independently authored fixture expects two accepted orders totaling 2600 pence. Vanilla ends with eight physical tubs after one receiving event; strawberry has one physical tub and four pending; chocolate stays at twelve. These are fixture expectations, not a business forecast. Its Telegram and model exchanges are simulated. A successful day does not establish a live handset exchange or long-running uptime.

Keep building with Prof Rod

Found this material through a colleague, classroom or shared download? Get the complete book at profrod.ai/book and join the Prof Rod learner community. Bring one result, one question or one failure you learned from. Share this resource with another learner and keep its source links with it so they can find the full course and future updates.

SOURCE PROVENANCE

Derived from github.com/profrodai/sovereign-agent/blob/5b825f3a58461bce8aba1173acf0b6d175c44b64/book/textbook/profrod-sovereign-agent-textbook-checkpoints.md