Skip to content
TakoAPI
Back to skills

codex-conductor

by shalomobongo

Coding Agents & IDEs
1.4kDownloads reported by clawskills.sh
9Stars
0Views
0 likesSign in to like this skill
—Not yet rated
Your rating

Brief

Methodical end-to-end software delivery orchestrator for Codex CLI with dual project modes (greenfield for new.

Description

What This Skill Does

An end-to-end software delivery orchestrator that coordinates Codex CLI or alternative coding agents through a strict stage-gated workflow. Supports greenfield (new builds) and brownfield (existing systems) project modes, each with autonomous or gated execution. No code is written without a written, approved spec with testable acceptance criteria.

Unlike one-shot code generators, it enforces a written spec with testable acceptance criteria before any implementation begins, preventing scope creep and requirement guessing across the full delivery lifecycle.

When to Use It

  • Building a new app from scratch with structured delivery gates and automated doc updates
  • Onboarding an existing codebase for incremental modernization with a risk register and rollback plan
  • Running gated delivery where each stage pauses for manual approval before the next begins
  • Tracking multi-task progress with a live dashboard showing gate state, blockers, and completion percentage
  • Generating per-task spec-driven prompts for Codex, Claude, or other coding agents one task at a time

Example Workflow

INPUT

User asks: Build a REST API with authentication from scratch using Codex CLI

AGENT
  1. 1

    Asks user to select project mode (greenfield) and execution mode (gated), then asks which coding agent to use

  2. 2

    Runs the planning questionnaire to capture mission, user journeys, v1 scope, stack preference, and definition of done

  3. 3

    Creates a written spec with acceptance criteria and waits for user approval before any code is written

  4. 4

    Initializes project docs, AGENTS.md, and .orchestrator/status.json via init_project_docs.py

  5. 5

    Generates gate-specific Codex prompts and processes tasks one at a time through gates G0 to G7, running validation commands and recording evidence in docs/validation-log.md after each task

OUTPUT

Delivered REST API with gate history in .orchestrator/status.json, test evidence in docs/test-results.md, and updated progress, change-log, and traceability docs

Requirements

  • At least one supported coding agent installed: Codex CLI (requires OpenAI account), Claude Code, opencode, or pi

Installation

clawhub install shalomobongo/codex-conductor

Run the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.

Links