jh941213/codex-lattice · Archived

plan-memory-hierarchy

Create backend/frontend layered `.md` memory during Plan mode. Use when request includes requirements capture, memory bootstrap, or plan preparation with backend/frontend separation.

First seen Aug 25, 2026

Installation

$ npx skills add jh941213/codex-lattice --skill plan-memory-hierarchy

Stronger alternatives

This repository is archived — consider an actively maintained alternative.

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from jh941213/codex-lattice · top by installs.

npx skills add jh941213/codex-lattice

Browse all from jh941213/codex-lattice

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 19
License LICENSE
Default branch main
Open issues 0
Status Archived

Skill metadata

Parsed from SKILL.md frontmatter.

Declared agents codex

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 2,469 B
  • docs SUMMARY.md 211 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 1 installs

SKILL.md

Layered Project Memory

This skill creates and maintains structured markdown memory files at project start, then reuses them during planning and execution.

Runtime Rules

  1. Confirm the project root. If unclear, ask the user.
  2. If .plan-memory/ does not exist, run bootstrap script.
python3 scripts/bootstrap_plan_memory.py --root . --project-name "<project-name>"
  1. Read core memory files:

- .plan-memory/README.md - .plan-memory/product/requirements.md - .plan-memory/backend/requirements.md - .plan-memory/frontend/requirements.md - .plan-memory/shared/constraints.md

  1. Store new information in the most relevant layer file, not in a single aggregate file.
  2. Before final plan output, update .plan-memory/README.md summary index.

Use Cases

  • New project intake with requirements capture
  • Splitting work by backend/frontend scope
  • Long-running refactor/integration work requiring persistent context

스크립트 동작

Default directory tree

.plan-memory/
  README.md
  product/requirements.md
  product/acceptance-criteria.md
  backend/requirements.md
  backend/api-contracts.md
  backend/data-model.md
  frontend/requirements.md
  frontend/ux-flow.md
  frontend/state-management.md
  shared/constraints.md
  shared/glossary.md
  execution/current-plan.md
  execution/open-questions.md

Key Rules

  • Existing files are preserved by default; do not overwrite unless --force is passed.
  • For edits, ask for user confirmation before deleting/replacing existing docs.
  • For new projects, fill execution/current-plan.md and product/acceptance-criteria.md first.

Maintenance Rules

  • Keep backend and frontend assumptions separate.
  • Put cross-cutting rules (security/performance/compliance) in shared/constraints.md.
  • Keep the summary index in README.md in sync with added layers/files.
  • End each planning pass by writing plan/risk/acceptance updates into execution/current-plan.md.

스크립트 위치

  • /Users/kdb/.codex/skills/plan-memory-hierarchy/scripts/bootstrapplanmemory.py

Failure Handling

  • Stop and ask for confirmation if root is unclear.
  • On filesystem errors, report exact failed path and error message.