8 views
This guide focuses on Capstone Build Week Checklist for production systems, with practical trade-offs for reliability, security, and cost.
Capstone Build Week Checklist
You have a spec (Day 98). This is the build-week gate list — security, traces, budgets, dual control, docs — before anyone schedules a demo.⚡ TL;DR: Do not demo what you cannot kill, trace, budget, or roll back. Walk this checklist daily. Red items block Day 100.
Daily board
| Area | Gate | Owner |
|---|---|---|
| Auth | Service identity only; no god keys | Platform |
| Tools | Allowlist + dry-run default | Squad |
| Dual control | Mutate requires second human | Squad |
| Traces | OTel from UI → gateway → model → tool | Platform |
| Budgets | Hard reject at cap | Platform |
| Prompts | Pinned version in gateway | Squad |
| Eval | CI green on golden set | Squad |
| Kill switch | Flag tested in staging | Both |
| Docs | README + runbook + ADR | Squad |
| Privacy | Redaction on logs/memory | Both |
bash
# ✅ Pre-demo smoke
curl -f "$GATEWAY/health"
curl -f "$GATEWAY/v1/complete" -H "Authorization: Bearer $TOK" -d @fixtures/golden1.json
# expect: trace id header, cost dims logged, dry-run tools only
Red-flag stops
- Mutate tool reachable without dual control
- No tenant filter on retrieval
- Prompt edited live in prod console
- Eval suite skipped “for speed”
- Kill switch never flipped in staging
Failure modes
Green CI with empty golden set. Require minimum N cases and a coverage note (injection, throttle, happy path). Staging that uses prod IAM “temporarily” — block demo if so.Closing checklist
- [ ] All table gates green or waived in writing
- [ ] Staging demo rehearsal recorded
- [ ] Rollback exercised once
- [ ] Cost dashboard shows the feature tag
- [ ] On-call briefed on kill switch names
Series navigation
Day 98: Capstone Spec: Pick One Production Problem · Day 100: Capstone Ship: Demo, Postmortem, Next 90 Days Last updated September 11, 2026 Related: Day 97: Build vs Buy the Agent Runtime; Designing a URL Shortener That Handles 100 Million Requests Per Day; Lambda Extensions API: Build Custom Monitoring Without Touching Application Code.Watch: 100 Days of AI on YouTube
Short videos from CheatCoders, one AI engineering topic per day.
Most viewed
- Python Decorators Explained: From Simple Wrappers to Production Patterns
- Claude Code vs Cursor vs GitHub Copilot: Choose by Workflow, and Set Up One Repo That Works With All Three
- AI Agent Frameworks in 2025: LangGraph vs CrewAI vs AutoGen vs Raw API
- Python String Methods: Every str Method With Real Production Examples
- Css Animation Performance: Performance-Optimized Motion for Modern UIs
Newly added
- AWS Systems Manager Parameter Store: Hierarchical Runtime Config for Coding Agents
- Amazon Cognito Identity Pools: Federate Coding-Agent Tool Callers Without Long-Lived Keys
- Amazon Athena for Coding Agents: Query Tool Audit Logs and Eval Traces on S3 Without a Warehouse
- AWS Batch for Coding Agents: Parallel Overnight PR Reviews and Long-Running Jobs Without Lambda Timeouts
- Amazon Bedrock Prompt Management: Version and A/B Test Coding-Agent System Prompts
Deep-dive PDF
Get the expanded guide for this post — extra diagrams-style checklists, failure modes, and a production walkthrough. Free when you subscribe to CheatCoders.
Already subscribed? or open the subscribe page.
Discover more from CheatCoders
Subscribe to get the latest posts sent to your email.
