| name | scrolls-help |
| description | Shows a crisp, example-driven reference for the whole scrolls skill family (/scrolls-setup, /scrolls-update, /scrolls-hide, /scrolls-unhide) — what each command does, their shared -p/-t/-l/-r/-u flags, and common usage examples (new project, monorepo, running from a subdirectory, toggling hidden/visible). Use this whenever the user runs /scrolls-help, or asks what the scrolls commands do, how to use them, what flags they take, which one they need, or seems unsure about docs/.scrolls setup — even if they only name one of the four commands rather than 'scrolls' generally. Supports -e/--online to render it as a styled page served on an OS-assigned localhost port instead of chat text, with light/dark and colorize/plain toggles. Works on macOS, Linux, and Windows (bash or PowerShell). |
| license | MIT |
| compatibility | bash (macOS/Linux/WSL) or PowerShell 7+ (pwsh) or PowerShell (older versions: powershell.exe); Python 3.9+ (stdlib only, no dependencies) |
| disable-model-invocation | true |
| metadata | [{"name":"scrolls-help","type":"skill","author":"sugatoray","version":"2.1.0","source_url":"https://github.com/sugatoray/aiskills/tree/master/skills/scrolls/scrolls-help"}] |
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
