Repository authority
•••Set instruction order, path scope, approval gates, and stop conditions before an agent changes a working tree.
Self-paced coding-agent operations
Use this course when a coding agent can change real code and you need a repeatable way to constrain, verify, release, and recover that work.
Curriculum
Each module ends with a working artifact. The workbook joins them into one capstone release.
Set instruction order, path scope, approval gates, and stop conditions before an agent changes a working tree.
Translate a request into observable acceptance, a change budget, and reversible implementation slices.
Match each release claim to risk-weighted checks and preserve evidence another operator can reproduce.
Classify changed state, set observable rollback triggers, and rehearse the path to a known-good outcome.
Review provenance, capabilities, dependencies, and re-vetting triggers without mistaking a scanner for a verdict.
Ship a signed immutable change, run a canary, preserve rollback authority, and measure activation.
Before agents
For people who use a model most days and are not sure the work holds up. Six modules with a lab each, a two-week workbook that runs your own project through all six, and a field guide. $59, self-paced, private reader and PDFs.
Work With AI · $5901
Four-Job Routing Matrix · 3 lessons
02
Verification Ladder · 3 lessons
03
Claim-Source Ledger · 3 lessons
04
Three-Pass De-Slop Sweep · 3 lessons
05
Explain-Predict-Break-Rebuild Loop · 3 lessons
06
Design Constraint Block · 3 lessons
What ships
Learn the method, practise the decisions, and keep the solo Agent OS Kit.
Skip the lessons and install the editable controls directly.
The kit alone is €149, so the full method — six modules, six labs, and the capstone — effectively adds €150. Live agentic-engineering cohorts in this space run $499 to $1,999 per seat; this is the same control loop as a written standard you keep.
Read the six-gate control guide