sdcorejs-subagent-driven-dev

Coordinate parallel subagent execution for approved multi-unit coding plans.

2|Updated Apr 18, 2026
One-click install
npx skills add https://github.com/sdcorejs/sdcorejs-agent --skill sdcorejs-subagent-driven-dev
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sdcorejs-subagent-driven-dev
Source: https://github.com/sdcorejs/sdcorejs-agent/tree/main/plugin/skills/sdcorejs-subagent-driven-dev
Command: npx skills add https://github.com/sdcorejs/sdcorejs-agent --skill sdcorejs-subagent-driven-dev

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Prevents parallel work from fragmenting into duplicate scaffolds and incompatible implementations by enforcing a structured, verifiable delegation workflow across subagents.

Core Features & Use Cases

  • Plan-based decomposition: Splits execution only from an already approved plan, ensuring units have independent file targets, isolated context, and isolation-verifiable tests.
  • Per-unit briefing protocol: Gives each subagent a zero-history, self-contained brief with explicit deliverables, verification commands, and explicit out-of-scope boundaries.
  • Controlled dispatch + wave handling: Fans out to multiple Agent calls in parallel (capped at 5) with wave sequencing for larger batches.
  • Partial-failure surfacing + synthesis: Requires honest reporting of ✅/⚠️/❌/🟡 outcomes, then performs conflict scans, convention drift detection, and redundant-code checks before a consolidated handoff.
  • Global verification gates: Runs build, lint, full tests, and smoke checks after all units complete to catch seam-level issues between delegated outputs.

Use cases include batch CRUD across multiple entities in one module, multi-screen generation, multi-stack/reference-doc batches, and any 3+ independent sub-tasks during 07-write-code orchestration (Angular Portal, NestJS, Next.js), including bilingual (VI/EN) reporting with permission codes in English.

Quick Start

Delegate execution by instructing the AI tool to run sdcorejs-subagent-driven-dev when the approved plan contains 3+ independent dispatch units during 07-write-code.

Frequently Asked Questions about sdcorejs-subagent-driven-dev

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent merge conflicts when delegating parallel coding tasks to subagents?

Preventing merge conflicts during parallel coding tasks requires enforcing strict unit decomposition with independent file targets and isolated per-agent briefings. This Skill runs conflict scans and convention drift detection before final handoff to ensure seamless integration.

What is the best way to orchestrate batch code generation across multiple independent modules?

Orchestrating batch code generation is best achieved through controlled parallel dispatch with wave sequencing. This Skill caps concurrent agent calls at five and provides each unit with a self-contained briefing to manage large batches across Angular Portal, NestJS, and Next.js.

Can I use subagent delegation for generating multiple screens in Next.js?

Yes, you can use subagent delegation for multi-screen generation in Next.js. The Skill supports delegating three or more independent work units during code generation phases, handling everything from initial dispatch to final global build and smoke verification.

How does partial-failure reporting work when parallel subagent execution fails?

Partial-failure reporting for parallel subagent execution requires honest surfacing of success, warning, error, or pending outcomes. The Skill mandates this status reporting before performing redundant-code checks and synthesizing a consolidated handoff.

Do I need an approved plan before starting subagent-driven development?

Yes, an approved plan is required before starting subagent-driven development. The Skill splits execution only from an already approved plan to ensure generated work units have independent file targets, isolated context, and isolation-verifiable tests.

What are the limitations of using parallel dispatch for large SDLC orchestration tasks?

A limitation of parallel dispatch for SDLC orchestration is that work units must be strictly independent to avoid merge chaos. Additionally, concurrent agent dispatch is capped at five, requiring wave sequencing for larger batches beyond that threshold.