feature-delivery

BEFORE building any feature/module; avoids reimplementing existing code or merging dead branches

When coordinating feature development with model-selection, telemetry, and integrated deployment.

Landing specialized sub-agents on specific models (opus/composer/codex) without turn-by-turn driving.

Full-cycle agile feature delivery: requirements → architecture → Confluence docs → Jira tasks → implementation → metrics.

Building features with test-first discipline across multiple implementation rounds.

workflow automation in specialized problem domain; check artifact name and description.

Agile teams converting informal feature requests into structured specs without heavyweight process overhead.

Orchestrate multi-phase evolving-builder-demo tasks with structured verification and recovery.

When features are complete and need clear, developer-actionable documentation

Implementing single features completely autonomously with deterministic TDD, review, and merge — reducing human context-switching.

Iterative code review → fix loops up to 3 rounds, hard-capped.

Starting a new feature from scratch; ensures design, planning, review, test, perf, docs all happen.

Orchestrating feature development by creating implementation-ready specs, committing to main, and creating corresponding feature branches with synchronized naming.

Implementing production features end-to-end with idiomatic code and comprehensive tests.

hen convenient |

Use when agents need to generate test cases, manage feature workflows or scope bugs from stories.

Stress-testing feature plans against dialect complexity and NonTerminal dispatch.

Providing subagent-level automation for feature implementation tasks.

When you have a detailed plan and need to implement code changes step-by-step with validation at each stage.

When you have a detailed plan and need to implement code changes step-by-step with validation at each stage.

Automating or structuring a commonly needed workflow.