@evolver/adapter-codex

OpenAI Codex CLI executor adapter for the Evolver framework.

Installation

pnpm add @evolver/adapter-codex

Usage

import { CodexExecutor, SkillConverter } from "@evolver/adapter-codex";

const executor = new CodexExecutor({
  name:        "codex",
  command:     "codex",
  skillsPath:  "./skills",
  skillFormat: "markdown",
  timeout:     60_000,
  concurrency: 2,
});

// Execute tasks (implements @evolver/core Executor interface)
const results = await executor.run(program, tasks);

// Convert skills to AGENTS.md format
const agentsMd = SkillConverter.mergeAgentsMd(skills);

API

CodexExecutor

Executes tasks via the Codex CLI (--quiet --approval-mode full-auto). Converts skills to a single AGENTS.md file, deploys to a temp directory, and runs tasks with configurable concurrency.

new CodexExecutor(config: AdapterConfig)
executor.run(program: Program, tasks: Task[]): Promise<ExecutionResult[]>

Security: uses execFile (no shell invocation) to prevent injection.

Built-in scorers: exact-match (default) and fuzzy.

SkillConverter

Converts SKILL.md format to Codex's AGENTS.md format. Each skill becomes a ## heading section.

SkillConverter.toAgentsMd(skill: Skill): string
SkillConverter.mergeAgentsMd(skills: Skill[]): string

Configuration

Uses AdapterConfig from @evolver/core:

| Option | Type | Description | |--------|------|-------------| | name | string | Adapter identifier | | command | string | CLI command (e.g. "codex") | | skillsPath | string | Path to skills directory | | skillFormat | "markdown" \| "json" \| "yaml" | Skill file format | | timeout | number | Per-task timeout in milliseconds | | concurrency | number | Max parallel task executions |

Testing

pnpm test