Skip to content
Back to skills

Roadmap

ASecurity

Read-only view of the execution roadmap. Lists every epic with its per-status tallies, nests the still-open rows beneath their epic, collapses wholly-finished epics into one rollup line, and closes by naming the next planned task in file order. Four verbs — `list` (the reader's view), `tasks` and `epics` (filterable row and epic queries), `next` (the first planned task). Invoke any time to see what is left to build. Not a workflow phase; never writes; never starts or commits work.

  • 14 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 4, 2026
ai-agentsnode

Works with

  • terminal
  • cli
  • mcp

Security analysis

A100/100

Pro scans all 4 files and shows the line behind each finding

Scanned October 4, 2026

npx -y skills add friedbotstudio/baseline --skill roadmap --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Roadmap?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Roadmap
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/friedbotstudio-roadmap/badge)](https://www.skillsdirectory.com/skills/friedbotstudio-roadmap)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: roadmap
owner: baseline
description: Read-only view of the execution roadmap. Lists every epic with its per-status tallies, nests the still-open rows beneath their epic, collapses wholly-finished epics into one rollup line, and closes by naming the next planned task in file order. Four verbs — `list` (the reader's view), `tasks` and `epics` (filterable row and epic queries), `next` (the first planned task). Invoke any time to see what is left to build. Not a workflow phase; never writes; never starts or commits work.
disable-model-invocation: true
---

# roadmap — what is left, and what is next

A read-only reader over the execution plan named by `project.json → roadmap.path`, in the family of `standup` and `audit-baseline`: it reads state and reports, it never mutates. It is **not** a workflow phase — it does not enter the Track Guard ordering and never blocks a commit.

`standup` answers six questions and the roadmap is one of them, so it bounds the roadmap section and drops to tallies on a large plan. This command answers only that one question, so it never collapses open rows at any plan size. What collapses instead is the finished work — that is the bulk, and it carries no pickup.

## The verbs

```
node .claude/skills/roadmap/cli.mjs list  [--all] [--epic N] [--json] [--root <dir>]
node .claude/skills/roadmap/cli.mjs tasks [--epic N] [--status done|in-progress|planned] [--json]
node .claude/skills/roadmap/cli.mjs epics [--json]
node .claude/skills/roadmap/cli.mjs next  [--json]
```

| Verb | Answers |
|---|---|
| `list` | The whole plan as a reader sees it — header, rollup, open rows nested under their epic, and the next pickup. |
| `tasks` | Individual rows, filtered by `--epic` and/or `--status`. One line per row. |
| `epics` | One line per epic with its status and its done / in-progress / planned tally. |
| `next` | The first planned row in file order, plus the epic it belongs to. |

Every verb takes `--json` and prints the raw projection instead of the rendering. A missing plan file exits 2 with the path it looked for; a malformed filter exits 1 with the usage error.

## What `list` prints

```
Roadmap — docs/roadmap-execution-plan.md
13 epics · 78 rows · 66 done, 0 in progress, 12 planned

✅ Epics 1-7, 10, 12  (53 rows, all done)

Epic 8  ⬜  Codebugger explanation trace  (0/3 done)
  ⬜ A  Runtime-witness rule and the mcp-debugger declaration
  ⬜ B  The /codebugger session and the explanation trace

Next planned: Epic 8  A  Runtime-witness rule and the mcp-debugger declaration
```

- **The rollup** folds every wholly-done epic into one line, compressing consecutive numbers into ranges. A lone finished epic reads `Epic 4`, never `4-4`.
- **Open rows** are planned and in-progress rows. Every one of them renders, however many there are.
- **`--all`** expands the rollup: each epic gets its own header and its done rows render too.
- **`--epic N`** scopes the view to one epic. Naming an epic is itself a request to see it, so a scoped view never collapses.

Every rendered line is whitespace-collapsed, clipped to 96 characters, and stripped of C0/C1 control characters by `.claude/skills/lib/terminal-text.mjs`. Roadmap titles are repository-controlled content on its way to a terminal, and a row title in this repo already runs past 1000 characters.

## Ordering

`next` and the `Next planned:` line both mean **first in file order** — the first planned row reading the plan top to bottom. Neither resolves a dependency graph; sequencing the work is `sprint-planner`'s job, not this reader's.

Files in this skill

  • SKILL.md3.5 KB
  • cli.mjs4.2 KB
  • parse.mjs5.3 KB
  • render.mjs4.4 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.