> ## Documentation Index > Fetch the complete documentation index at: https://docs.getunbound.ai/llms.txt > Use this file to discover all available pages before exploring further. # Discovery — Summary > Security posture, risk distribution and the users risk leaderboard | | | | ----------- | ----------------------------- | | **URL** | `/ai-tools-discovery/summary` | | **Sidebar** | AI Tools Discovery → Summary | | **Roles** | Admin · Analytics Viewer | Summary is the front page of AI Tools Discovery: how exposed your fleet is, what your developers are running, and who carries the most risk. ## What's on the page ### Security posture | Element | Shows | | ------------------------------ | ---------------------------------- | | **Security Posture Score** | Your fleet's overall standing | | **Security Posture Breakdown** | What is contributing to that score | The score runs 0–100 and higher is healthier: 76 and above reads Low, 51–75 Medium, 26–50 High, 25 and below Critical. Five exposures pull it down: | Exposure | What raises it | | --------------------- | ------------------------------------------------ | | **Autonomy exposure** | Developers whose tools may act without asking | | **Guardrail gaps** | Developers with neither deny rules nor a sandbox | | **Config file risk** | Config files rated Critical or High | | **Supply chain risk** | MCP servers from unverified publishers | | **Shadow AI** | Tools nobody has sanctioned | A fleet can read healthy overall while one of the five is bad, so the breakdown is the useful half: it names the exposure dragging the score down — the one to fix first. ### Distribution charts | Chart | Shows | | --------------------- | ----------------------------------------------- | | **Risk Distribution** | How risk is spread across your fleet | | **Tool Distribution** | Which AI tools are in use and how widely | | **Autonomy Levels** | How much independence the discovered tools have | ### Users Risk Leaderboard | Column | Contents | | ---------------- | ---------------------------------------------- | | **User** | The developer | | **Risk Score** | Their current score · sortable | | **Risk Factors** | What is driving it | | **Tools** | The AI tools found on their devices | | **Autonomy** | How autonomous their setup allows agents to be | | **Last Sync** | When their device last reported | A device that has not reported for more than three days is marked **Stale**, with the tooltip *No report in 3+ days*. ## Filters and controls | Filter | Values | | ------------------ | --------------------------------------------- | | **Risk Level** | Critical · High · Medium · Low | | **Autonomy Level** | Full Auto · Auto Edit · Ask First · Read Only | | **Tool** | The AI tools in your fleet | **Search users** narrows the leaderboard on any of name, username, system user, email or device ID. **Stale** keeps it to devices that have not reported in three days or more. **Risk Score** and **Last Sync** both sort, in either direction — selecting the same header again reverses it. The distribution charts double as filters: selecting a slice of **Risk Distribution**, **Tool Distribution** or **Autonomy Levels** applies it to the leaderboard, and selecting the same slice again clears it. ### The view is a link Your leaderboard filters and sort are written into the address as you set them, so a filtered view is a link anyone can open on exactly what you were looking at: | In the address | Means | | --------------------------- | -------------------------------------------------- | | `?risk_level=critical,high` | Those risk bands | | `?autonomy_level=full_auto` | That autonomy level | | `?tool=` | One or more tools, comma separated | | `?filter=stale` | **Stale** on | | `?sort_by=last_sync` | Sorted by **Last Sync** rather than **Risk Score** | | `?risk_sort=asc` | Lowest first | The search box is not carried, so a shared link opens the same filters with the search cleared. ## What you can do | Action | Result | | ---------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Sanction MCP Servers** | Marks servers as approved, bringing them into your sanctioned set | | **Sanction agents** | The same for AI tools | | **Export Users Risk Leaderboard** | CSV of the leaderboard, under the filters and search currently applied. Seven columns are included by default and six more can be ticked, among them **Permission Risk** and **Config File Risk**, which the table itself shows only on hover | | Select a developer | Opens their per-user view, below | | **Start setup** on the empty state | Opens the product's setup page at `/configure` | ### Per-developer views Selecting a row on the leaderboard opens that device's own view at `/ai-tools-discovery/users`, with the device already selected — its tools, MCP servers, config files, skills and risk breakdown in one place. ## Messages you may see | Message | What it means | | ---------------------------------------------------- | -------------------------------------------------------------------- | | `Could not start sync` | The sync request was not accepted. Try again shortly | | `Failed to load incidents` | The incidents view could not load. Refresh to retry | | `Could not read a complete engineer list for export` | The export could not gather the full list. Retry for a complete file | ## Related | Page | For | | ----------------------------------------------------------- | ---------------------------------- | | [Discovery — Inventory](/dashboard/discovery/inventory) | The catalogue behind these figures | | [Discovery — Permissions](/dashboard/discovery/permissions) | How much autonomy those tools have | | [Configure](/dashboard/configure) | Getting devices reporting |