Skip to content
Apothem
Comparison

Apothem vs raw Codex config

Hand-maintaining Codex's AGENTS.md, hooks, skills, and agents versus installing them from Apothem's shared profile.

Codex reads user instructions from AGENTS.md, loads hooks from hooks.json or config.toml, discovers custom agents under agents/, and discovers skills from the shared Codex skills root. Apothem's codex adapter installs those surfaces from the shared profile while leaving operator-owned config.toml settings alone.

Effort comparison

TaskRaw hand-configApothem-managed
Author AGENTS.mdHand-edit user instructionsInstalled from the Codex bootstrap template
Author hooksHand-write hooks.json or TOML hook tablesInstalled to ~/.codex/hooks.json with Apothem hook helpers under ~/.codex/hooks/
Author custom agentsHand-write TOML files under ~/.codex/agents/Converted from Apothem Markdown agents
Author reusable commandsHand-build skills one directory at a timeWrapped as ~/.agents/skills/*/SKILL.md
Keep rules aligned with other harnessesManual, per toolAutomatic on apothem update
Verify the config is correctManual inspectionapothem verify --harness codex

What Apothem manages

The codex adapter installs AGENTS.md, hooks.json, hook helper files, converted custom agents, and generated skills. It does not overwrite ~/.codex/config.toml; model, sandbox, approval, MCP, and other runtime preferences remain operator-owned.

Migration path

  1. Run apothem profile init to create the shared profile if you have not yet.
  2. Move the rules from your existing AGENTS.md into the shared corpus.
  3. Keep existing config.toml runtime settings in Codex unless you choose to manage them manually elsewhere.
  4. Run apothem install --harness codex to install the Codex-native surfaces.
  5. Run apothem verify --harness codex to confirm the files landed.
  6. Use apothem update --harness codex for every later change.

On this page