Skip to content
Skills Library
Management & product · Planning

lantern-plan

/lantern-plan

Plan a change in a project that has a settled `.lantern/` map. Writes a structured plan file under `.lantern/plans/` that the Lantern app can read - goal, scope, impact on areas, features and files computed from the map, ordered stages with items and acceptance checks, the open questions of each stage, risks - and has the user approve it section by section. Use whenever the user asks to "plan", "plan this feature", "lantern plan", "zaplanuj", "make a plan before we delegate", "break this into stages", "what would this touch" or "review the plan", in a project with a `.lantern/map`. Refuses to

Repository
lantern ↗
Licence
MIT
Last repository change
13.11.2026

Assessment and security

90/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content25 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
25
Scripts
22
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 13.11.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What to check before installing
  • 22 scripts in this package. Review them before installing, as Claude may execute them.

Scripts: scripts/plan-approval.mjs, scripts/plan-check.mjs, scripts/plan-edit.mjs, scripts/plan-flow.mjs, scripts/plan-format.mjs, scripts/plan-gate.mjs, scripts/plan-git.mjs, scripts/plan-handoff.mjs, scripts/plan-impact.mjs, scripts/plan-io.mjs, scripts/plan-links.mjs, scripts/plan-map.mjs, scripts/plan-messages.mjs, scripts/plan-model.mjs, scripts/plan-parse.mjs, scripts/plan-run.mjs, scripts/plan-shape

What it does and when to use it

Plan a change in a project that has a settled `.lantern/` map. Writes a structured plan file under `.lantern/plans/` that the Lantern app can read - goal, scope, impact on areas, features and files computed from the map, ordered stages with items and acceptance checks, the open questions of each stage, risks - and has the user approve it section by section. Use whenever the user asks to "plan", "plan this feature", "lantern plan", "zaplanuj", "make a plan before we delegate", "break this into stages", "what would this touch" or "review the plan", in a project with a `.lantern/map`. Refuses to

Triggers from the description

planplan this featurelantern planzaplanujmake a plan before we delegatebreak this into stageswhat would this touchreview the plan
Original author description

Plan a change in a project that has a settled `.lantern/` map. Writes a structured plan file under `.lantern/plans/` that the Lantern app can read - goal, scope, impact on areas, features and files computed from the map, ordered stages with items and acceptance checks, the open questions of each stage, risks - and has the user approve it section by section. Use whenever the user asks to "plan", "plan this feature", "lantern plan", "zaplanuj", "make a plan before we delegate", "break this into stages", "what would this touch" or "review the plan", in a project with a `.lantern/map`. Refuses to

How it works

  1. Read before each step
  2. Language
  3. Process
  4. Commands
  5. Rules
Original SKILL.md headings · 83 lines, 841 words
SKILL.md preview
---
name: lantern-plan
description: Plan a change in a project that has a settled `.lantern/` map. Writes a structured plan file under `.lantern/plans/` that the Lantern app can read - goal, scope, impact on areas, features and files computed from the map, ordered stages with items and acceptance checks, the open questions of each stage, risks - and has the user approve it section by section. Use whenever the user asks to "plan", "plan this feature", "lantern plan", "zaplanuj", "make a plan before we delegate", "break this into stages", "what would this touch" or "review the plan", in a project with a `.lantern/map`. Refuses to start while the map is not settled. This is the entry skill - the gate, the process, and which small `lantern-plan-*` skill to read at each step. Not for building the map (`lantern-setup`) and not for doing the work (`lantern-delegate`).
---

# Lantern plan

A plan here is a file, not a document: `.lantern/plans/<slug>.plan`, line-based, tied to the map, committed
next to it. Lantern reads it, the user approves it section by section, and `lantern-delegate` executes it
item by item. Item progress is never written in the plan; it comes from the pull request records that name
the item. Format: `../lantern-setup/references/format.md`, "Plans".

`$P` is this skill's folder, `$L` is `lantern-setup`'s. Run `node $P/scripts/plan.mjs <command>` from the project root. Every script
is here; the other `lantern-plan-*` skills point back at it.

## Read before each step

Read the skill named for the step, from disk, in full, before you do the step. `plan.mjs skills` lists
any that are missing.

| step | what | read first |
|---|---|---|
| 1 | gate and request | `lantern-plan` |
| 2 | impact from the map | `lantern-plan-impact` |
| 3 | stages, items and questions | `lantern-plan-stages` |
| 4 | write and check the plan file | `lantern-plan-stages` |
| 5 | review section by section | `lantern-plan-review` |
| 6 | approve the plan | `lantern-plan-review` |
| 7 | hand off to lantern-delegate | `lantern-plan-handoff` |

## Language

Write every plan text (goal, scope, impact reasons, stage and item titles, acceptance, questions, risks) in the
map's primary language: `node $L/scripts/lang.mjs .` prints its code (`pl`, `en`, ...), and it comes from the
map itself, so nothing needs configuring. Plan keywords, ids and file paths stay as they are.

## Process

1. **Gate and request.** Run `plan.mjs gate`. Closed means stop: say in one line what failed. A map that is
   not settled: tell the user to set it up or refresh it first (`lantern-setup`). A missing skill: install
   with `pnpm run install-skills -- <project>`. Never plan on a map you could not validate.
   Then read the request and the map (`.lantern/features`, `.lantern/areas`, the texts in `lang/`). List
   only what you cannot find out and cannot sensibly default: ask it in **one message**, numbered, at most
   five, each with the answer you would assume. Not a questionnaire. A question that only blocks one stage
   waits for that stage (step 3).
2. **Impact.** Compute it from the map, do not guess it. Write the `#impact` section.
3. **Stages.** Split the work into ordered stages of items, each with an acceptance check; put each open
   question in the stage it blocks.
4. **Write and check.** Pick the slug from the goal: a few lowercase words joined by hyphens, at most 40
   characters (`plan-on-the-map`). `plan.mjs new`, then `plan.mjs add`, or edit the file; `plan.mjs check <slug>`
   must print OK.
5. **Review.** Show one section, get the user's answer, record the approval, show the next.
6. **Approve.** When every section is approved, record the plan-level approval. Commit the file.
7. **Hand off.** An approved plan goes to `lantern-delegate` as `Plan: <slug>`.

## Commands

| command | does |
|---|---|
| `gate` | map settled (`validate.mjs --settled`) and planning skills installed; exit 1 closed |
| `skills` | which planning skills resolve, and what to read at each step |
| `new <slug> --title <t> --goal <g>` | starts the file |
| `impact --file/--feature/--area ...` | dependents, dependencies, counts, areas and features |
| `add <slug> <kind> ...` | adds a goal, scope, impact, stage, item, touch, question, answer, explain, flow or risk line (impact and touch take the kinds area, feature, file and new; a touch may belong to a question; flow takes `--flow-file`) |
| `check <slug>\|--all` | shape, names against the map, approvals; exit 1 on errors |
| `show <slug> [<section>]` | the review queue, or one section as written |
| `approve <slug> <section\|plan> --by <who>` | records an approval |
| `verified <slug> <item> --by <who>` | the one status the plan stores |
| `status <slug>` | item status from the PR records, and what comes next |
| `handoff <slug>` | the items to delegate now; exit 1 unless approved and clean |

Every command takes `--json` and `--root <dir>`.

## Rules

- **Short lines, no prose blocks.** A goal is at most three lines, a scope bullet one line; `check` rejects
  more. If a section wants a paragraph, it is two sections or a question.
- **Names come from the map.** Features, areas and files in the plan must exist in `.lantern/`; copy them
  from `plan.mjs impact`. `check` errors on an unknown one and warns on a file deleted since.
- **Questions belong to a stage.** Never one list at the end.
- **Answers arrive pasted from Lantern.** The user answers in the app, which writes nothing; you record the
  answers, any `explain` or `flow` it asks for, and any new question (`lantern-plan-stages`).
- **Approval is the user's.** Only `approve` after the user said yes to what you showed. You never approve
  for them, and an edit after approval makes that section stale: show it again.
- **The plan never records progress** except `#verified`. Do not tick items by hand.
- **Scratch to `.ai/tmp/`.** Write only `.lantern/plans/`.
Files in this package (25)
  • README.md 2 KB
  • SKILL.md 6 KB
  • scripts/messages.json 31 KB
  • scripts/plan-approval.mjs 2 KB
  • scripts/plan-check.mjs 6 KB
  • scripts/plan-edit.mjs 5 KB
  • scripts/plan-flow.mjs 4 KB
  • scripts/plan-format.mjs 3 KB
  • scripts/plan-gate.mjs 2 KB
  • scripts/plan-git.mjs 6 KB
  • scripts/plan-handoff.mjs 3 KB
  • scripts/plan-impact.mjs 6 KB
  • scripts/plan-io.mjs 2 KB
  • scripts/plan-links.mjs 2 KB
  • scripts/plan-map.mjs 2 KB
  • scripts/plan-messages.mjs 1 KB
  • scripts/plan-model.mjs 1 KB
  • scripts/plan-parse.mjs 11 KB
  • scripts/plan-run.mjs 1 KB
  • scripts/plan-shape.mjs 3 KB
  • scripts/plan-status.mjs 4 KB
  • scripts/plan-write.mjs 3 KB
  • scripts/plan.mjs 12 KB
  • scripts/planning-skills.mjs 2 KB
  • scripts/selftest.mjs 8 KB

Versions

The database has no other skills named lantern-plan.

Database snapshot: 13.11.2026 · package version e73d3d1066c8 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 90lantern-plan-review
    Show a Lantern plan to the user one section at a time and record each approval, then the plan-level approval. Use at steps 5 and 6 of `lantern-plan`, and whenever the user asks to "show me the plan", "review the plan", "approve the plan", "go through the plan with me", "pokaż…
  • 100lantern-plan-stages
    Split a planned change into ordered stages of items, give every item an acceptance check and the features and files it touches, put each open question in the stage it blocks, then write and check the `.plan` file with `plan.mjs add` and `plan.mjs check`. Use at steps 3 and 4 of…
  • 90create-with-lantern
    Use Lantern's application MCP to create, draw, compose, revise, organize, inspect, or review a creator's comics through the domain capabilities Lantern currently exposes. Use for comic work in Lantern; do not use for developing the Lantern source repository.
  • 91serve
    Start the lantern web UI for a map, where the user answers, comments, approves and ends their turn. Use for `/lantern:serve [slug]`.
  • 96map
    Chart a new lantern map, a round-based brainstorming DAG that leads the user through the fog of a big idea over many sessions. Use for `/lantern:map <idea>`.
  • 90flutter-plan
    Turns an app idea, a design or a feature request into screens, user flows and journeys (Mermaid diagrams) and an ordered build plan of small slices with backend, service, native and design needs, acceptance checks and effort ranges. Use when asked to plan, map flows or journeys,…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.