# Learn > Use when the user wants to learn the sn-toolkit plugin itself, including what is new or what shipped recently in it. An interactive guided tour of the toolkit's own capabilities. Do NOT use for ServiceNow platform/API questions (GlideRecord, ACLs, business rules -- that is sn-toolkit:docs), and do NOT use when the user is operating on a specific record or instance. - **Type:** Skill - **Install:** `agentstack add skill-chrisp28103-sn-toolkit-learn` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [chrisp28103](https://agentstack.voostack.com/s/chrisp28103) - **Installs:** 0 - **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [chrisp28103](https://github.com/chrisp28103) - **Source:** https://github.com/chrisp28103/sn-toolkit/tree/main/sn-toolkit/skills/learn ## Install ```sh agentstack add skill-chrisp28103-sn-toolkit-learn ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About ## What this skill is A guided tour of the **sn-toolkit plugin's own** capabilities -- not a ServiceNow platform tutorial. It exists because the toolkit is large (30+ commands, plus skills, subagents, hooks) and both new and returning users lose track of what is in it. The catalog is **generated live** from the installed plugin via `sn-learn.ps1`, so it is always current -- a newly shipped command appears in the tour with no edits here. Your job is to *teach* it well: route to the right mode, keep it interactive, and never dump the whole thing at once. All calls use `powershell.exe -NoProfile -File sn-learn.ps1 `. If it is not on PATH, use the full path `${CLAUDE_PLUGIN_ROOT}/bin/sn-learn.ps1`. ## Step 0 -- disambiguate if the trigger was loose (cheap insurance) This skill can auto-fire. If the request is ambiguous -- e.g. "what can I do here", "how do I do X" where X looks like a ServiceNow platform topic, or anything that reads like work on a live record -- do NOT launch a tour. Ask ONE line first: > "Want a tour of the sn-toolkit plugin, or did you mean [the ServiceNow thing they > said]? For platform questions I'll use the docs lookup instead." A misfire then costs one sentence. Only proceed into a mode once intent is clear. ## Step 1 -- pick the mode | User signal | Mode | Section | |---|---|---| | New, "where do I start", "how do I use this", no specifics | `new` (default) | Step 2 | | "show me everything", "what commands are there", "full list" | `all` | Step 3 | | "what's new", "what shipped", "what changed since I last used it" | `whats-new` | Step 4 | | Names one command/skill ("how does /pull work") | `` | Step 5 | If `$ARGUMENTS` already names a mode or a command, skip straight to it. ## Step 2 -- mode `new`: walk the golden path (curated spine) New users do not need all 30 commands -- they need the **core dev loop**, in order. Teach this as the backbone, one step at a time, pausing to let them act: 1. `/sn-toolkit:new-project ` -- bootstrap the workspace + the sn-scriptsync instance bridge (the one piece everything else rides on). 2. `/sn-toolkit:creds` -- store instance credentials (DPAPI, gitignored). 3. `/sn-toolkit:start` -- verify the round-trip (server up + browser connected). 4. `/sn-toolkit:pull ` -- read an existing record's script/config. 5. `/sn-toolkit:update` (or `/sn-toolkit:widget`) -- make a change. `update` writes directly via the Agent API and verifies persistence itself (script fields included); nothing more is needed for most edits. 6. `/sn-toolkit:sync-push` -- *only if you opted into the local file-sync workflow* (e.g. editing widget files locally): flush + confirm it landed, error-checked. 7. `/sn-toolkit:end` -- wrap up, update trackers for next session. Before sending them into commands, confirm the prerequisites exist (sn-scriptsync extension + SN Utils browser tab) -- without them every command times out. Point them at the README "Prerequisites" section if unsure. End by offering: *"Want the full catalog (`all`), or to go deeper on any one of these?"* ## Step 3 -- mode `all`: the live catalog, grouped ``` sn-learn.ps1 index ``` This prints every command, skill, and agent with a one-line summary. Do NOT paste it raw -- regroup it for the user using the categories below (the same ones the README uses), then ask which group they want to drill into. - **Connect / inventory**: start, end, creds, list, new-project, instance - **Read**: pull, export, review, audit - **Write**: create, update, widget, component, sync-push - **Session context**: switch - **Visual debugging**: inspect, attach - **Planning**: refine-prompt - **Documentation / deliverables**: spec, workshop, workbook, worddoc, docs, docs-setup, docs-sync - **Operations**: monitor, diagnose, catalog-test - **Experimental (agent teams, opt-in)**: diagnose-team, compare-team, compare - **Subagents**: SN Explorer, SN Reviewer, SN Platform Admin If `index` lists a command NOT in any group above, surface it under **"Newer / not yet categorized"** and note it is a recent addition -- that is the signal to update this section. (This is the only maintenance this skill needs as the toolkit grows.) ## Step 4 -- mode `whats-new`: the returning-user lane ``` sn-learn.ps1 whats-new [baseline-version] ``` Omit the version for just the latest release; pass a version (e.g. `1.18.0`) to show everything since. Read it from the CHANGELOG and **summarize the user-facing deltas** -- new commands, behavior changes that affect how they work -- not the internal refactors. Lead with anything new they can try now. If the CHANGELOG is not found in the install, the script prints the GitHub link; hand that to the user. ## Step 5 -- mode ``: one command, with a worked example ``` sn-learn.ps1 show ``` Returns that item's full description, its argument hint, and a capped body peek. Use it to explain: what the command does, when to reach for it, a concrete example invocation, and which commands it pairs with (e.g. `pull` -> `update` for a direct edit; or, if you want local version control, `pull` -> edit the synced file -> `sync-push`). Keep it to a short, runnable walk-through, not a recital of the body. ## Constraints - **Interactive, not a wall.** One mode, one workflow at a time. Always end by offering the next step. The whole point is that the README is too dense to absorb at once. - **Never re-dump.** If you already showed the catalog this session, reference it; do not reprint it. - **Stay in your lane.** Platform/API questions -> `/sn-toolkit:docs`. This skill teaches the *toolkit*, not ServiceNow. - **ASCII only** per sn-toolkit rule #1. The script already sanitizes its output. - **Read-only.** This skill never runs SN commands on the user's behalf -- it teaches the user to run them. Do not invoke `/sn-toolkit:*` work commands from inside the tour. ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [chrisp28103](https://github.com/chrisp28103) - **Source:** [chrisp28103/sn-toolkit](https://github.com/chrisp28103/sn-toolkit) - **License:** MIT Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** no - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/skill-chrisp28103-sn-toolkit-learn - Seller: https://agentstack.voostack.com/s/chrisp28103 - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.