Scrolls help

This skill's job is to answer clearly and get out of the way — not to explore the user's codebase, not to check whether scrolls are actually set up here, and not to modify anything in the user's project.

Scope note: This skill never writes to user code, user files, or anything outside its own directory while carrying out a /scrolls-help request. Keeping its own references/HELP.md current is a separate, unrelated maintenance task for people maintaining this skill — see meta/MAINTAINERS.md — not something done while answering a user. For the full security posture (file-system scope, the local -e/--online server's network exposure and lifecycle), see meta/SECURITY.md — reference material, not something to act on while answering a request.

Cross-platform

-e/--online's launcher ships in two forms: open_help.sh (bash — macOS, Linux, or Windows with Git Bash/WSL) and open_help.ps1 (PowerShell 7+ — Windows, or macOS/Linux with pwsh installed); serve_help.py, the actual server both of them launch, needs no porting at all — it's pure stdlib Python, already cross-platform, and both launchers call it the same way (resolving python3 then falling back to python, since Windows commonly has only the latter on PATH). Pick the launcher by what's actually available: try bash --version; if that succeeds, use .sh; otherwise use .ps1 via pwsh (preferred — install from https://aka.ms/powershell if missing) or, only if pwsh genuinely isn't available, the built-in Windows PowerShell powershell.exe (untested against that older version; pwsh is what this was written and verified against).

Options

Read the invocation text for an optional -e / --online — instead of (or in addition to, if it's ambiguous which the user wants) answering in chat, render references/HELP.md as a styled HTML page and serve it locally. Everything else about interpreting the request (bare invocation vs. a specific question) works exactly the same regardless of this flag; -e only changes the output channel.

Steps

1. Read references/HELP.md — that's the maintained, canonical content. Present it rather than reconstructing an explanation from memory of how these skills work: flag semantics here have changed across iterations (e.g. -r used to mean "repo root," now means "recurse"), and the reference file is the single source of truth that gets updated when that happens.

2. -e/--online given: run bash <skill-dir>/scripts/open_help.sh or pwsh <skill-dir>/scripts/open_help.ps1 (see "Cross-platform" above). It launches a small stdlib-only Python server (no dependencies to install) that renders HELP.md into a clean, self-contained HTML page and binds it to 127.0.0.1 on a port the OS assigns (never all interfaces; this is a local reference viewer, not something to expose on the network). The script waits for the server to confirm it's actually listening before printing anything, then prints the URL followed by the process's PID. Report that URL to the user as a clickable link — it keeps running after this skill finishes, so the link stays open — and mention that it shuts itself down automatically after 30 minutes idle or 2 hours total (whichever comes first), and can be stopped sooner with <script> --stop <port> (or --stop --all to stop every instance) — no need to hunt down and kill a PID by hand. A fresh invocation with -e starts another server on a new port rather than reusing one, which is fine but worth knowing if several accumulate over a long session; each stops on its own even if never told to. If the script's own browser-opening attempt didn't visibly do anything (e.g. a headless/remote environment with no display), that's expected — the URL is still valid and the report to the user is what matters.

   The page has a small toolbar (top-right) with two independent toggles: light/dark (follows the system preference by default; the button forces either explicitly, persisted via localStorage) and colorize/plain (code blocks get a small GitHub-syntax-style token palette by default — commands, flags, and comments each colored distinctly, with separate light and dark values, in the spirit of a Pygments/pymdown-extensions theme; the button strips that back to plain text, also persisted). Mention both toggles are there when reporting the URL — they're not obvious from the link alone.

3. Bare /scrolls-help, or an open-ended question ("what are the scrolls commands," "how does this work"), without -e: present the whole document, as markdown chat output. This is an answer, not a deliverable — don't write it to a file or publish it as an artifact unless the user separately asks for that.
4. A specific question (one command, one flag, one scenario like "how do I do this in a monorepo," or a command name mentioned on its own): lead with the directly relevant part of the doc, answered concisely, and mention that the rest is available via a bare /scrolls-help — don't dump the whole reference regardless of what was actually asked. This applies whether or not -e was also given — -e changes where the full doc goes, not whether a targeted question still gets a targeted answer first.
5. Keep the reference file's own tone and formatting when you present it (in chat or via the rendered page) — it's already written to be crisp; don't editorialize, pad, or re-explain what it already says clearly.

Examples

Example invocations of /scrolls-help itself are in assets/EXAMPLES.md. Read that file only when you need it, for example to check how -e or --stop behaves. It is not the content to present: references/HELP.md is. For worked examples of another scrolls command, point the user at that skill's own assets/EXAMPLES.md.

Development

See meta/MAINTAINERS.md for how to keep references/HELP.md current and how to run the tests/ regression suite. Neither is read as part of carrying out a user's /scrolls-help request — don't act on meta/MAINTAINERS.md or tests/ while answering one.