CloakBrowser MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| CLOAK_MCP_SCREENSHOT_DIR | No | Screenshot output directory path | |
| CLOAK_MCP_DEFAULT_CDP_URL | No | Default CDP endpoint URL | |
| CLOAK_MCP_DEFAULT_DISPLAY_MODE | No | Default display mode when not provided (headless, virtual, or cdp) |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| browser_startC | Start a CloakBrowser session. Default mode is direct headless Linux browsing. |
| browser_navigateD | Navigate the session page to a URL. |
| browser_clickC | Click an element by CSS selector. |
| browser_typeB | Type text into an element by CSS selector. |
| browser_evaluateC | Evaluate JavaScript in the current page. |
| browser_snapshotC | Return page URL, title, and visible text. |
| browser_screenshotC | Capture a PNG screenshot and return its filesystem path. |
| browser_closeB | Close a browser session and release resources. |
| browser_wait_for_selectorC | Wait for an element selector to reach a page state. |
| browser_pressC | Press a keyboard key while targeting an element. |
| browser_hoverC | Hover over an element by CSS selector. |
| browser_select_optionC | Select an option value in a select element. |
| browser_get_textB | Return visible text from the page or a selector. |
| browser_get_attributeC | Return an element attribute value. |
| browser_get_linksB | Return links from the page or a selector. |
| browser_scrollC | Scroll the active page by pixel deltas. |
| browser_reloadC | Reload the active page. |
| browser_go_backC | Navigate the active page back in history. |
| browser_go_forwardC | Navigate the active page forward in history. |
| browser_get_cookiesC | Return browser cookies. |
| browser_set_cookiesD | Set browser cookies. |
| browser_clear_cookiesC | Clear browser cookies. |
| browser_get_storage_stateC | Return browser storage state. |
| browser_save_storage_stateC | Save browser storage state to a file. |
| browser_new_pageC | Open a new page in the session. |
| browser_list_pagesC | List pages in the session. |
| browser_switch_pageC | Switch the active page by page ID. |
| browser_close_pageB | Close a page in the session. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 28 tools
Each tool has a clearly distinct purpose, targeting specific browser actions like navigation, clicking, typing, or cookie management. Potential overlaps like browser_get_text and browser_snapshot are differentiated by scope (snapshot includes URL and title). No two tools are ambiguous.
All tools share the 'browser_' prefix, but the verb style varies: some are simple verbs (browser_click, browser_close) while others follow verb_noun (browser_clear_cookies). Minor inconsistency but still predictable and readable.
28 tools is on the higher side but appropriate for a comprehensive browser automation server covering navigation, interaction, extraction, state management, and cookies. No tools seem redundant or out of scope.
The tool set covers a wide range of browser automation needs: navigation, clicking, typing, text extraction, screenshots, JavaScript evaluation, cookies, storage, page management, and waiting. Missing advanced features like file upload or network interception, but core workflows are well covered.