// SCAN REPORT

Wraith

17 tools. 11 can modify or destroy data without limits.

11 write tools that can modify data. Rate limits recommended.

Last updated: 2 Jul 2026

11 can modify or destroy data

6 read-only

17 tools total

Community server · catalogue entry verified 02/07/2026

How to control Wraith ↓

TOOL MAP

What Wraith exposes to your agents

Read (6) Write / Execute (11) Destructive / Financial (0)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

High Risk

The most dangerous Wraith tools

Execute · High navigate Navigation in a browser automation context is an Execute operation—it triggers external effects (loading web pages, executing JavaScript, accessing remote servers) whose consequences depend on the URL argument. Execute · High click Clicking UI elements in a browser is an Execute-category action because the effects depend entirely on what is clicked — it can trigger purchases, deletions, form submissions, or other operations. Execute · Medium press_key This tool triggers browser keyboard input actions (key presses and chords) that can submit forms, trigger shortcuts, or interact with web page elements. Execute · Medium scroll Scrolling is a browser interaction that triggers page-level effects (e.g., lazy-loading content, infinite scroll pagination, triggering JavaScript scroll events).

11 of Wraith's 17 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Wraith

PolicyLayer is an MCP gateway — it sits between your AI agents and Wraith, and nothing reaches the server without passing your rules. These are the rules we recommend:

Rate limit write operations

{
  "close_all_sessions": {
    "limits": [
      {
        "counter": "close_all_sessions_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "extract": {
    "limits": [
      {
        "counter": "extract_per_minute",
        "window": "minute",
        "max": 60,
        "scope": "grant"
      }
    ]
  }
}

Controls API costs and prevents retry loops from exhausting upstream rate limits.

Create a free account and register Wraith — nothing to install.
Add these rules — paste them, or build them visually. Tune the limits to your setup.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

ENFORCE POLICY ON WRAITH →

Instant setup, no code required.

FULL CATALOGUE

All 17 Wraith tools

EXECUTE 7 tools

Execute browse browse Execute click Click the element with the given index (from snapshot). Returns a new snapshot. Execute navigate navigate Execute press_key Press a key or chord (e.g. "Enter", "Control+a"), then return a new snapshot. Execute scroll Scroll the page, then return a new snapshot. Execute tabs Manage browser tabs within a persistent session. Execute type_text Type text into the element with the given index. Returns a new snapshot.

WRITE 4 tools

Write close_all_sessions Close all active persistent browser sessions. Write close_session Close a persistent browser session. Write pdf Render a webpage to PDF and return it as a base64-encoded PDF. Write save_storage_state save_storage_state

READ 6 tools

Read extract extract Read get_content Return the current page as clean markdown (no LLM, no screenshot). Read list_downloads list_downloads Read list_sessions List active persistent browser session IDs. Read screenshot Take a screenshot of a webpage and return it as base64-encoded PNG. Read snapshot Return the current page's interactive elements as [index] <tag> text.

EXPLORE

Related servers

Other MCP servers with similar tools — same risk classification, starter policies for each.

FAQ

Questions about Wraith

How do I prevent bulk modifications through Wraith? +

The Wraith server has 4 write tools including close_all_sessions, close_session, pdf. Set a rate limit in your policy -- for example, 10 calls per hour prevents an agent from making more than 10 modifications per hour. PolicyLayer enforces this at the gateway, before calls reach Wraith.

How many tools does the Wraith MCP server expose? +

17 tools across 3 categories: Execute, Read, Write. 6 are read-only. 11 can modify, create, or delete data.

How do I enforce a policy on Wraith? +

Register the Wraith MCP server in PolicyLayer, apply the suggested rules above (adjust the limits to your use case), and point your AI client at the PolicyLayer proxy URL instead of the server directly. Your agents keep the same tools; PolicyLayer evaluates every call against policy before it executes. Nothing to install, live in minutes.

Enforce policy on every Wraith tool call.

Deterministic rules across all 17 Wraith tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON WRAITH →

Instant setup, no code required.

17 Wraith tools catalogued and risk-classified — across an index of 43,000+ MCP servers.