Omniscio

Tools View (browse installed CLIs + linked skills)

The Tools view is a dedicated virtual project that lists every CLI Omniscio integrates with, across ten categories, with each row showing install status and its detected version. Selecting a tool opens a detail pane with the description, an Install button for auto-installable tools, and chips for the skills that depend on it, which jump to that skill in the Skills view.

On this page

What it is

The Tools view is a dedicated virtual project that lists every CLI Omniscio integrates with — currently 30 tools across ten categories (AI Agents, Runtimes, Git, Cloud, Containers, Media, Browser, Skill CLIs, System, Security). Every one is a public tool any user can get; a CLI that exists only on one person's machine is never part of this list, and instead appears (italic, under Installed) on that one machine once the list's Scan now button finds it on the PATH. Each tool row shows install status and, when installed, its detected version. Selecting a tool opens a detail pane with the description, install status chip, an Install button for auto-installable tools (or a View Instructions link for system-level ones like Node.js), and a Used by skills row with chips for the skills installed on this machine that depend on that tool. Clicking a skill chip jumps straight to the Skills view with that skill selected. Settings → Connected Tools (the update-focused panel) is kept, not replaced — it still owns auto-update behaviour, the pending-updates queue, and developer overrides. The Tools view is the discovery surface ("what's available and what needs a skill?"); Toolchain Settings is the control surface ("how do updates roll out?").

Playwright MCP and Superpowers are not shown in the Tools view because they aren't CLI binaries — they're Claude Code concepts (an MCP server registered in ~/.claude.json, and a plugin from a marketplace). They have their own onboarding step and post-onboarding nudge banner. See mandatory-claude-extensions.md.

Where to find it

Click Tools (wrench icon) in the left sidebar — it appears as a virtual project next to Skills, Recipes, Gmail, etc. From Settings → Connected Tools you can also click the See all tools → link next to the "Connected Tools" description.

How it behaves

How to use it

  1. Open it. Click Tools (wrench icon) in the left sidebar — it appears as a virtual project next to Skills, Recipes, Gmail, etc. From Settings → Connected Tools you can also click the See all tools → link next to the "Connected Tools" description.
  2. Browse by category. The sub-sidebar groups tools under category headers (AI Agents, Runtimes, Git, …) in a stable order. Each row shows the tool name plus an Installed (emerald) or Error (red) badge when a check result is cached; uncached rows show no badge. Type in the search box to filter by name, id, or description (case-insensitive).
  3. Inspect a tool. Click any row. The detail pane shows the tool name, category + type chips, description, and a status chip: Installed vX.Y.Z (click to copy the version to clipboard) or Not installed. Error status surfaces the probe error inline with an alert icon.
  4. Install a tool. For auto-installable tools (autoInstallable: true), the detail pane shows an Install button that calls TOOLCHAIN_INSTALL_TOOL. While the install runs the button shows a spinner (isInstalling state from toolchain-update-store), and on success the status chip flips to the new version. Non-auto-installable tools (e.g. Node.js — OS-level) show a View Instructions link to their docs site instead. A tool that reads installed but whose own check reports its setup unfinished keeps this same button, labelled Finish setup, and shows no Update while it is offered — the state a tool carries when it was already on the computer but the part the app still owes it (such as a skill agents learn it from) was never placed.
  5. Jump to linked skills. When a tool has linkedSkillIds authored, a Used by skills row appears with one chip per linked skill that is installed on this machine — a linked skill you do not have is not shown, and the row disappears when none are installed (prefixes like user: / plugin: are stripped for display). Click a chip to navigate to the Skills view with that skill pre-selected — handy when a skill's install guide tells you to grab ffmpeg first.
  6. Where the recommended set comes from. First-run setup shows this same catalog as one short consent list before anything is installed: the tools Omniscio itself needs are ticked, the optional ones sit beside them, and each row says where its install comes from (winget, npm, the vendor's own installer). Only the rows you leave ticked are installed, and a tool you decline stays declined — nothing re-installs it behind your back later. Opt-in tools are deliberately kept out of that sweep (OpenCode, an alternative coding agent, and portless among them) and are installed when you want them, from their own row here, from the Running Apps panel, or with the one-click setup button in Settings. A couple of tools instead arrive the moment a flow needs them: the GitHub CLI is installed just-in-time by the GitHub-connect, diff and PR flows rather than being a first-run requirement.
  • toolchain-update-v3.md — how tool updates (silent vs deferred) are handled from Settings → Connected Tools
  • mandatory-claude-extensions.md — Playwright MCP + Superpowers plugin (Claude-Code-side extensions, installed in parallel with tools during onboarding)
  • use-skills.md — the Skills view that linked-skill chips navigate into

Last verified 2026-10-16