**See exactly which instructions your coding agent loads for any file, and what it silently drops.** `axm explain` reads the CLAUDE.md files, AGENTS.md files, rules and imports in a repo, and shows what Claude Code and Codex each load for a given file, in order, with the reason for every one.

Instructions for coding agents now live in many files: a global CLAUDE.md, the project's AGENTS.md, nested files per directory, path-scoped rules and imports. Each harness loads them by its own rules, and none of them tells you when something is skipped. A `CLAUDE.local.md` makes Claude Code stop reading your AGENTS.md. A rule with broken frontmatter loads for every file instead of the ones it names. Codex stops adding files once they reach 32 KiB. You find out when the agent ignores a rule you know you wrote.

**Status:** planning. There is nothing to install yet. See [ROADMAP.md](ROADMAP.md).

How it works

The output below is the planned v0.1 output for a small repo. Its `CLAUDE.md` imports a missing file, doesn't import `AGENTS.md`, and sits next to a rule scoped to `src/api/**`.

$ axm explain src/api/orders/OrderService.cs --harness all

CLAUDE CODE // launched at repo root
  01  ~/.claude/CLAUDE.md          user                at launch
  02  CLAUDE.md                    project             at launch
  03  .claude/rules/backend.md     paths: src/api/**   when the file is read
  --  AGENTS.md                    DROPPED  CLAUDE.md exists and doesn't import it
  --  src/api/AGENTS.md            DROPPED  CLAUDE.md exists and doesn't import it
  --  docs/testing.md              DROPPED  imported by CLAUDE.md:12, file is missing

CODEX // launched at repo root
  01  ~/.codex/AGENTS.md           global              at launch
  02  AGENTS.md                    project