Atomizer — Development agent for Claude Code
Converts PlanOutline (+ optional RepoFindings) into atomic tasks with DoD and dependencies (DAG).
How to install Atomizer
Installs to ~/.claude/agents/bohewu-agents_pipeline-atomizer.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/bohewu/agents_pipeline/HEAD/agents/atomizer.md -o ~/.claude/agents/bohewu-agents_pipeline-atomizer.md Restart Claude Code, or start a new session, for it to be picked up.
What Atomizer does
name: atomizer description: Converts PlanOutline (+ optional RepoFindings) into atomic tasks with DoD and dependencies (DAG). kind: subagent
ROLE
Produce atomic TaskList (DAG). Each task must be independently verifiable.
INPUT RULES
- Required planning inputs are
PlanOutlineand any explicit scope constraints in the handoff. - If
DevSpecis present, use it to keep tasks behavior-oriented and traceable. - When
DevSpecis present, each task MUST include a non-emptytrace_ids
Alternatives in Development
- Gitnexus Docs Dod Reviewer — GitNexus docs and Definition-of-Done reviewer 45.8k ★
- Executioner — Optional Sanhedrin fallback verifier 632 ★
- Helm Conductor — Use this agent to orchestrate engineering work cycles — decide what to work on next, sequence stories and feat 523 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Reordering Agent
Анализ зависимостей между секциями детальной архитектуры и пересортировка в правильный порядок реализации. DAG
Change Scoper
Break work into small, trunk-sized tasks with binary DoD and safe rollback.
Feat Implement
Execute implementation plan tasks systematically with optional batch mode
Workflow Planning Dependency Grapher
Analyzes authored tasks to establish internal dependencies and priorities. Invoked by workflow-planning-proces
Coordinator Agent
Manages complex workflows, coordinates dependencies between tasks, and ensures proper execution order across d
Ralph Conductor
Orchestre les sessions Ralph Wiggum v2.0 avec validation DoD adaptative
Related Skills
Ralph Output Adversarial Review
Adversarially reviews the artifacts produced by a RALPH run (code/tests/UI/config) against repo policy (testin
Parallel Dispatcher
Dispatch multiple task-workers in parallel based on DAG dependencies.
Team Tasks
Multi-agent pipeline coordination: Linear, DAG, and Debate modes for AI agent orchestration