Babysitter:help

babysitter:help

Help and documentation system for the babysitter Codex CLI plugin.

No Arguments — Show Welcome

If no arguments provided, display this welcome:

Babysitter for Codex CLI — Orchestration Plugin

Primary Commands:
  /babysitter:call        Start an orchestration run (interactive)
  /babysitter:resume      Resume an existing run
  /babysitter:yolo        Start a run (non-interactive, no breakpoints)
  /babysitter:plan        Plan a workflow without executing
  /babysitter:forever     Start a never-ending periodic run

Secondary Commands:
  /babysitter:doctor      Diagnose run health (10 checks)
  /babysitter:retrospect  Analyze a run and improve future processes
  /babysitter:model       Set/view model routing policy
  /babysitter:issue       Start workflow from GitHub issue
  /babysitter:team-install Install team-pinned setup from lockfile
  /babysitter:assimilate  Assimilate external methodology
  /babysitter:user-install   Set up babysitter for yourself
  /babysitter:project-install  Onboard a project
  /babysitter:observe     Launch observer dashboard

Type /babysitter:help <command> for detailed help on a specific command.

With Arguments — Show Details

If an argument is provided:

1. Command help: Read the SKILL.md for that command from .codex/skills/babysitter/<name>/SKILL.md and display its content
2. Process help: Read the process .js file from .a5c/processes/<name>.js and describe it
   • Also search bundled upstream process library at upstream/babysitter/skills/babysit/process
3. Skill/agent help: Use wrapper discovery helpers; if invoking SDK CLI directly use babysitter skill:discover --plugin-root "$CODEX_PLUGIN_ROOT" --json
4. Methodology help: Search the process library for matching methodology

Use the skill-loader module to resolve command names:

const { getSkillContent } = require('./.codex/skill-loader');
const content = getSkillContent(args);