Practical DCP Flow use cases
Eighteen end-to-end scenarios combining Discovery, architecture/stack/cloud, Quality Gates, Git/multi-agent work, deploy/release, incidents, upgrades, and the website + external CRM case with EspoCRM as a contextual ADOPTED option.
Eighteen end-to-end scenarios combining Discovery, architecture/stack/cloud, Quality Gates, Git/multi-agent work, deploy/release, incidents, upgrades, and the website + external CRM case with EspoCRM as a contextual ADOPTED option.
Practical use cases show how multiple DCP Flow layers combine in real situations. They are not rigid recipes; they teach how to select artifacts, gates, and tools according to problem and risk.
End-to-end scenarios
The eighteen scenarios cover new product discovery, existing-project adoption, hosting/cloud decisions, multi-agent work, gate defects, permission/data changes, artifact validation, migrations, incidents, and upgrades.
- Discovery and ADOPT/CONFIGURE/INTEGRATE/EXTEND/BUILD.
- Architecture, stack, and cloud.
- QA, Git, and multi-agent work.
- Deploy, incidents, and upgrades.
Apply the methodology in context
Other scenarios cover tool/MCP evaluation, Kai Capability vs Authorization, mature products shifting toward support and upgrades, promotion of repeated lessons, and websites that integrate an external CRM without becoming the CRM themselves. In that scenario, EspoCRM is the ADOPTED lightweight CRM option, while selection remains contextual.
- Tool Catalog and MCP evaluation.
- Kai and Human Gates.
- Mature-product operations.
- Promotion of Lessons Learned.
Discovery and adoption in real scenarios
Use cases show how to approach a new product without assuming BUILD, how to reconstruct Project Truth in an existing repository, and how to choose shared hosting, VPS, or cloud by separating architectural requirements from provider preference.
- Capability Matrix before building.
- Audit before massive refactor.
- Architecture requirement ≠ provider preference.
- Proportional decision by context.
Multi-agent work, gates, and risk
Other scenarios show how to create Concurrency Contracts for two agents, classify Product/Gate/Harness Defects before changing correct architecture, and treat small changes with high impact —such as permissions or data exposure— according to risk rather than line count.
- Temporal ownership and Shared Interfaces.
- Git Conflict ≠ Functional Conflict.
- Gate defect should not distort the product.
- Change size ≠ risk level.
Artifacts, releases, incidents, and Kai
The Handbook applies Artifact/Package Gates when a ZIP may fail despite green source and CI; defines Migration Ordering and recovery for high-risk releases; structures Incident Response; evaluates tools before adoption; and separates Kai capability from authorization when actions may be destructive.
- Source correct ≠ artifact correct.
- Release with backup/recovery evidence.
- Stabilize → understand → recover → verify.
- Capability ≠ Authorization.
Contextual glossary＋
Practical scenario showing when and how to combine parts of DCP Flow.
Critical end-to-end journey that must work and produce verifiable evidence.
Explicit, verifiable manual procedure used when Kai cannot execute an action directly.
Checkpoint where evidence determines whether work may progress.