ワンクリックで
pipeline
Configurable pipeline orchestrator for sequencing stages
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
メニュー
Configurable pipeline orchestrator for sequencing stages
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Strategic planning with optional interview workflow
Coordinated parallel execution for complex multi-lane tasks
Ask Claude via local CLI and capture a reusable artifact
Ask Gemini via local CLI and capture a reusable artifact
Full autonomous execution from idea to working code
Fix build and TypeScript errors with minimal changes
| name | pipeline |
| description | Configurable pipeline orchestrator for sequencing stages |
$pipeline is the configurable pipeline orchestrator for oh-my-kimi. It sequences stages
through a uniform PipelineStage interface, with state persistence and resume support.
The canonical oh-my-kimi pipeline sequences:
RALPLAN (consensus planning) -> team-exec (Kimi Code CLI workers) -> ralph-verify (architect verification)
Pipeline parameters are configurable per run:
| Parameter | Default | Description |
|---|---|---|
maxRalphIterations | 10 | Ralph verification iteration ceiling |
workerCount | 2 | Number of Kimi Code CLI team workers |
agentType | executor | Agent type for team workers |
Every stage implements the PipelineStage interface:
interface PipelineStage {
readonly name: string;
run(ctx: StageContext): Promise<StageResult>;
canSkip?(ctx: StageContext): boolean;
}
Stages receive a StageContext with accumulated artifacts from prior stages and
return a StageResult with status, artifacts, and duration.
prd-*.md and test-spec-*.md planning artifacts already exist, and carries any deep-interview-*.md spec paths forward for traceability.Pipeline state persists via the ModeState system at .omk/state/pipeline-state.json.
The HUD renders pipeline phase automatically. Resume is supported from the last incomplete stage.
state_write({mode: "pipeline", active: true, current_phase: "stage:ralplan"})state_write({mode: "pipeline", current_phase: "stage:<name>"})state_write({mode: "pipeline", active: false, current_phase: "complete"})import {
runPipeline,
createAutopilotPipelineConfig,
createRalplanStage,
createTeamExecStage,
createRalphVerifyStage,
} from './pipeline/index.js';
const config = createAutopilotPipelineConfig('build feature X', {
stages: [
createRalplanStage(),
createTeamExecStage({ workerCount: 3, agentType: 'executor' }),
createRalphVerifyStage({ maxIterations: 15 }),
],
});
const result = await runPipeline(config);