project-memorylisted

# Project memory — taxonomy and tooling **Codex scope:** this skill documents and configures Claude Code memory. Codex may use the taxonomy or maintain an existing Claude setup, but `autoMemoryDirectory` and the migration scripts do not configure native Codex memory. Do not run `init` merely to make Codex remember something. Follow the consuming project's AGENTS.md memory instructions; native Codex memory setup is outside this skill's supported scope. Resolve `<plugin-root>` to the installed package directory containing this skill's `skills/` parent, using the loaded skill's absolute path. Do not assume a shell variable is set by the client. Substitute that absolute path in the commands below; both Claude Code and Codex can run them from the target repository. The **single consultable source** for where agent memory lives, how it moves, and how a repo is wired to hold it. The first half is the memory taxonomy standard a session applies; the second half is the two bundled scripts that realize its structural side. The system *design* behind it — storage mechanics, loading semantics, settings precedence — is the memory-standard technical design (`docs/design/memory-standard.md`); this content is the standard a session applies plus the as-built tools, and it links to that design rather than restating it. Three consumers read this identical content: a human/agent session (you, now), the repo-compliance-audit skill (checks structure against `references/checklist.md`), and the o