Agent based development

Skill RealDougEubanks/ClaudeMarketplace/skills/agent-based-development/skills/agent-based-development

Full async multi-agent development workflow: Planning → Design → Dev → Security/Tech Review loop with file-based handoffs and release-branch Git model.From its SKILL.md

Install

npx -y skills add RealDougEubanks/ClaudeMarketplace --skill agent-based-development

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.0 KB, 902 tokens by cl100k_base, as published. Nobody here has run it

Agent-Based Development — Async Multi-Agent Workflow

Run a full async multi-agent software development workflow using file-based artifact handoffs. Each invocation adopts a specific agent role and operates according to that role's responsibilities.

This is the orchestrator skill. Identify your single assigned role first (from the slash command used or the latest plan artifact), then follow that role's skill. Do not attempt to act as multiple roles in one invocation.

Framework Overview

This is an async, artifact-driven workflow. Agents do not communicate in real time — they communicate via JSON files in a handoffs/ directory. The flow is:

Planning → Design → Dev (Senior / Junior)
       ↓                        ↓
  Documentation ← Security + Tech Review
       ↓                        ↓
  Planning (triage) ← findings with severity

All agents must follow the Golden Rules (see CLAUDE.md or run /golden-rules).

Agent Roster

Role skills are namespaced by this plugin (e.g. /agent-based-development:abd-plan). Where your Claude Code resolves unambiguous bare names, the short form also works.

Agent	Skill	When to Include
Planning	abd-plan	Almost always — assigns tasks, triages reviews
Design	abd-design	When the project has a design phase; omit for simple/script-only work
Dev Senior	abd-dev-senior	When there is higher-complexity implementation work
Dev Junior	abd-dev-junior	For routine tasks; omit for small teams
Security	abd-security	Recommended for all projects
Tech Review	abd-review	Recommended for all projects
Documentation	abd-docs	Almost always — docs, ToDo, changelogs, assumptions
Planning Triage	abd-triage	After reviews land — rework assignments
Testing	abd-test	When the project has automated tests
DevOps	abd-devops	When the project has CI/CD or a formal release process
Status	abd-status	At any time — shows project-wide task and finding status

Active agents are defined in docs/agentRoster.md. If that file does not exist, Claude creates it at project start.

Shared References

Every role reads these before acting:

shared/envelope.md — handoff directory structure, envelope schema, artifact rules, prompt-injection guard, assumptions log, output format
shared/git-rules.md — release-branch Git model and responsibilities
shared/mvp-gate.md — readiness checks before marking any task complete
shared/project-start.md — bootstrap procedure and prompt examples

How to Use This Skill

Step 1 — Determine context. Use Glob to check if handoffs/ exists in the current working directory.

Step 2 — If no handoffs/ directory exists (project start): follow shared/project-start.md — adopt the Planning role and bootstrap the project.

Step 3 — If handoffs/ exists (ongoing project): read the latest artifact in handoffs/plans/ using Read and Glob, determine which role is assigned, and follow that role's skill from the roster table above.

Step 4 — Follow the artifact rules, assumptions log, and output format in shared/envelope.md, and the Git rules in shared/git-rules.md. Verify the MVP gate before marking any task complete.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most context ai engineering skills give in 902 tokens

Counted across 1,328 of the 2,349 authors here whose files we hold, read 2026-09-06

Dispatch a fresh subagent for each taskin 76 of 1328, across 59 files
Perform spec compliance review before code quality reviewin 44 of 1328, across 34 files
Dispatch a final code reviewer after all tasksin 38 of 1328, across 26 files
Answer subagent questions before allowing implementationin 36 of 1328, across 26 files
Use the least powerful model capable of the taskin 33 of 1328, across 26 files
Create a TodoWrite list for all tasksin 32 of 1328, across 22 files
Perform a task review after each implementationin 31 of 1328, across 24 files
Extract all tasks and context from the planin 29 of 1328, across 20 files
Provide full task text to subagentsin 28 of 1328, across 20 files
Use git worktrees for isolated workspacesin 25 of 1328, across 20 files
Specify the model explicitly when dispatching a subagentin 23 of 1328, across 18 files
Execute all tasks from the plan without stoppingin 21 of 1328, across 16 files

Said here and by no other author read

identify your single assigned role first
read the latest artifact in handoffs directory
follow the assigned role skill
verify the mvp gate before marking tasks complete
follow the golden rules
read shared project references before acting

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.