// Use before declaring work complete on business-critical phases — simulate a Senior practitioner (auth specialist, payment engineer, DBA, security auditor) reviewing the change with their failure-mode bias
[HINT] Download the complete skill directory including SKILL.md and all related files
| name | domain-expert-simulation |
| description | Use before declaring work complete on business-critical phases — simulate a Senior practitioner (auth specialist, payment engineer, DBA, security auditor) reviewing the change with their failure-mode bias |
Non-programmer cannot do code review themselves. They lack domain expertise to spot what's missing. Simulating a Senior practitioner mid-review compensates: Claude has seen the patterns AND the failure modes.
This is NOT performance theater ("look how thorough we are"). It's a structured search for what's wrong.
/gsd-ship on auth, payments, infra, security, or privacy changesYou are a Senior [DOMAIN] Engineer with 10 years experience at [COMPANY-PEER].
You have been on-call during [SPECIFIC-OUTAGE-TYPE] outages.
Review this change:
[paste code or PR diff]
Specifically check:
1. [Domain killer question 1]
2. [Domain killer question 2]
3. [Domain killer question 3]
Reject any concern as "edge case" only if it can NEVER happen in production.
Then fix what you find. Don't dismiss as "nitpick" — Senior X knows things you don't.
For high-stakes phases (auth, payments, breaking API):
Two layers catch different classes of issues:
Run both for security or money phases.
❌ Performance theater (asking simulation just to validate your design)
✅ Asking simulation to find what's wrong
❌ Cherry-picking (discarding answers that disagree)
✅ Treating disagreement as red flag worth investigating
❌ Single perspective on critical change
✅ Multiple domain experts on auth/payments (security + privacy + UX)
❌ Vague framing: "Review this change"
✅ Specific framing: "Review as Senior Payment Systems Engineer who has been on-call during Black Friday outage"
For these, regular code review or /audit code is enough.
components/domain-expert-simulation.md — full methodcomponents/supreme-council.md — multi-LLM external review (next layer)rules/non-programmer-safeguards.md — auto-trigger before shipcommands/audit.md — code-level audit (different layer)