Engineering Playbook
AIM
Define what we are aiming at before implementation begins.
Aiming is not bureaucracy. It is the cheapest place to discover bad assumptions, and the stage where a vague target gets made explicit.
What breaks without it
Without deliberate planning, assumptions hide until implementation — where correction is expensive and confidence is fake.
The deliverable
The Execution Contract
The agreement between the human who is accountable and whoever — or whatever — implements.
Planning becomes more important, not less, once implementation is cheap. The failure mode of AI-assisted delivery is rarely code that does not work. It is code that works correctly and implements the wrong thing.
If you aim at nothing, you will hit it. An execution contract exists to make the target explicit before anyone starts moving toward it — and to make that failure detectable before the work begins rather than after it merges.
It is reviewed once, at the level of strategy, rather than approving every individual action.
- Intent — the outcome being pursued, stated without prescribing the solution
- Scope — what is included, and explicitly what is not
- Constraints — architecture, security, data, performance and compatibility
- Risk — what could go wrong, and what makes this change hard to reverse
- Tests — how correctness will be demonstrated
- Acceptance criteria — testable conditions for done
- Execution plan — sequence, affected components and approval points
Practices
- Approve the plan, not each keystroke — strategy is the scalable control point
- Name at least one serious alternative that was considered and rejected
- Decide the test strategy before implementation, not after
- Flag irreversible decisions explicitly — they deserve more scrutiny