Methodology¶
The constraint-first approach to AI-assisted development.
The Principle¶
Write the rules before the code. Enforce them automatically. Stop at first failure.
This inverts the typical AI workflow:
- Spec first - define invariants, decisions, and architecture before implementation
- Gate before generate - every request passes through a guardian that checks alignment
- Stop at first failure - the gate chain halts immediately on any failure
- Escalate, don't guess - when the spec is silent, the system waits for a human decision
The Gate Chain¶
Every /wk-impl runs in order:
A failure at any step stops the chain. You never get code that violates the design.
Guardian Verdicts¶
| Verdict | Meaning |
|---|---|
| PASS | Proposal aligns with all invariants and ADRs. Proceed. |
| BLOCK | Violates a specific rule. Fix the proposal or the spec. |
| NEEDS-DECISION | Spec doesn't cover this. Record a decision first. |
Comparison with Ceremony-Heavy Approaches¶
| Aspect | Ceremony-heavy | Wolverine Kit |
|---|---|---|
| When roles run | Every feature | Phase-triggered |
| Enforcement | Guidelines | Machine verdicts |
| On ambiguity | Agent guesses | Stops, escalates |
| On violation | Retry and hope | Cites the rule |
Writing Good Constraints¶
- Be specific enough to check. "Code should be clean" is not an invariant. "No bare
print()outsideui.py" is. - State what's blocked, not what's encouraged.
- One decision per ADR. Keep them atomic so agents can cite them precisely.
- Update the spec when reality changes. A stale spec is worse than no spec.
- Fewer rules, strictly enforced. Ten invariants that always hold beat fifty that are sometimes checked.