Use gald3r with OpenAI Codex¶
OpenAI Codex — the codex CLI, an IDE integration, and a hosted app, all governed by a single schema-validated config.toml (model, sandbox/approval, MCP, hooks, agents, skill registration). The CLI is the full-extensibility surface; the IDE/app share the config plus an MCP settings panel.
Installing gald3r into an OpenAI Codex project writes the CRASH overlay this tool actually reads into .codex/ — the same task board, bug tracker, and gald3r <verb> CLI every other supported platform gets, wired the way OpenAI Codex itself expects it. Today that's commands native, rules native, skills native, hooks native, MCP native via the gald3r engine's MCP (L2). Cheapest full-parity install: ship gald3r's .agents/skills/g-skl-/SKILL.md (open standard) + an AGENTS.md — Codex loads both natively. Map g-agnt- to .codex/agents/.toml, wire g-hk-.py via hooks.json/[hooks] (10 events), declare MCP under [mcp_servers.], and fold g-rl- always-apply rules into AGENTS.md (hard guards optionally as execpolicy rules / PreToolUse hooks).
Once installed, the same .gald3r/TASKS.md task board, bug tracker, and full gald3r <verb> CLI work identically to every other supported platform — OpenAI Codex is one of 37 platforms gald3r reads and writes today. Commonly searched as Codex CLI, Codex.
Get started¶
gald3r setup
gald3r platform install codex