More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Tetherplane

    1

    Operate authorized Windows devices via Tetherplane MCP with scoped file, process, and browser controls.

    Free

    Security scanned
    Tetherplane

    Tetherplane

    Example session with this skill installed

    Check the health of my authorized Windows device through Tetherplane. Identify which tool namespaces and operations are available, check the connection state and local policy mode, and report any restrictions. Do not change files, execute commands, control the desktop, or interact with browser tabs.

    • Read your context and instructions
    • Compiled the tetherplane

    Tetherplane — Device Health & Capability Report

    Result: VERIFIED — Device responsive

    Read-only inspection performed November 5, 2026, using the connected Tetherplane MCP integration.

    1. Device Health

    CheckObserved result
    DeviceAuthorized Windows workstation (anonymized)
    Agent version0.1.0
    Operating systemWindows x86-64
    Agent processAlive
    ResponsivenessResponsive
    Connection transportAuthenticated
    Agent stateReady
    Local policybackground_only

    Assessment: The connected agent successfully responded to health and capability-discovery requests. Authentication and responsiveness are confirmed for this session.

    2. Available Tool Capabilities

    NamespaceReported availabilityNotes
    deviceAvailableStatus and capability discovery
    filesAvailableScoped filesystem operations
    processAvailableProcess execution and inspection, subject to policy
    batchAvailableGrouped operations with per-operation authorization
    desktopAvailableWindows semantic desktop provider; sensitive actions remain policy-controlled
    browserUnavailableBrowser provider not active on this device

    Additional native capabilities include search, job handling, and auditing where exposed by the connected implementation.

    Important: A provider reporting itself as available does not automatically authorize every operation it supports.

    3. Security and Permission Assessment

    • The workstation reports background_only policy.
    • No foreground-control permission was requested or exercised.
    • No filesystem writes or process executions were performed.
    • Browser automation was not attempted because the browser provider was unavailable.
    • Device-specific authorization still applies to all subsequent operations.

    The agent must evaluate the available operation, its authorization requirements, and its potential side effects before executing it.

    4. Completed Verification

    The following read-only MCP operations returned successfully:

    device.status — confirmed operating system, agent version, process liveness, responsiveness, transport state, and policy mode.

    device.capabilities — returned the device's advertised providers and operations.

    Outcome: Connectivity and capability discovery are verified for this authorized workstation. The report does not claim that browser automation, installation on a new machine, or arbitrary remote operations have been successfully completed.

    5. Recommended Next Step

    Choose an operation supported by the connected agent, confirm its local permission scope, and execute a small, authorized task with an independently observable result.

    For browser-related work, the unavailable provider must be configured and explicitly authorized separately. Do not bypass that restriction with another control method.


    This report is based on an actual read-only Tetherplane MCP observation, anonymized for public demonstration. It is not a production certification, customer onboarding acceptance test, or guarantee of universal compatibility.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Tetherplane — Developer Beta

    Tetherplane gives AI agents structured instructions for performing authorized computer operations through an existing, authenticated Tetherplane MCP connection.

    Designed for developers working with agent-driven workflows, it emphasizes permission-aware execution, local security policies, non-disruptive operation, and verifiable results.

    This free developer-beta skill requires the separately installed Tetherplane MCP runtime.

    Capabilities

    • Device operations: Check device connectivity, status, and available capabilities.
    • File operations: Read, write, and manage files within explicitly authorized directories.
    • Process execution: Run permitted commands and inspect their results.
    • Browser operations: Interact with supported browser sessions when the required browser integration and owner permissions are available.
    • Desktop automation: Use supported Windows UI operations, subject to local policy and human-activity protections.
    • Batch workflows: Coordinate supported operations while preserving execution boundaries and reporting outcomes.

    Requirements

    A separately installed, configured, and authenticated Tetherplane MCP connection is required.

    Tetherplane Operator does not install the Windows agent, grant operating-system permissions, or bypass device security controls.

    Developer Beta

    This is an early developer release. Features and availability vary by device, connection configuration, local policy, and owner authorization.

    Some operations may be unavailable in particular environments.

    Open source: https://github.com/muhamedsohaib/Tetherplane

    Publisher: ARCUS Labs

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.

    Tetherplane Computer Operations

    Secure, permission-aware computer operations for AI agents.

    Tetherplane Computer Operations is a skill that guides AI assistants through operating authorized Windows machines using the open-source Tetherplane Model Context Protocol (MCP) integration.

    Instead of giving an AI unrestricted desktop control, it starts by identifying the connected machine, discovering available capabilities, respecting device-local authorization, and verifying the outcome of permitted operations.

    What the skill helps your agent do

    • Discover: Check connection health, device capabilities, and supported operations before taking action.
    • Operate: Work with approved files, processes, and supported desktop or browser contexts.
    • Respect permissions: Preserve human-owned resources and use only locally authorized capabilities.
    • Verify: Inspect command results, returned handles, and relevant application state rather than assuming an action succeeded.
    • Recover: Diagnose disconnections, denied operations, timeouts, and uncertain outcomes without unsafe retries.

    How it works

    1. Install and connect the separately distributed Tetherplane MCP service.
    2. Authorize the device and relevant capabilities.
    3. Load this skill into a compatible AI-agent environment.
    4. Ask the agent to perform an authorized computer task.
    5. The agent follows the skill's instructions for discovery, execution, verification, and error handling.

    What makes it different

    Tetherplane prioritizes semantic operations, compact MCP responses, local policy enforcement, and non-disruptive background execution.

    Its six-tool Compact MCP interface groups device, file, process, browser, desktop, and batch operations without exposing a large collection of unrelated tools to the AI assistant.

    Requirements and limitations

    This is a developer-beta instructional skill, not a standalone remote desktop application. It does not include a Windows executable, automatically install the MCP server, create device permissions, or guarantee access to every supported provider.

    Browser and desktop capabilities are conditional on the connected device and its owner-approved configuration. Production readiness, clean-machine onboarding, and full multi-device acceptance are separate validation requirements.

    Open-source repository: https://github.com/muhamedsohaib/Tetherplane

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 10 days ago

    • Free with an Agensi account

    Software
    Tetherplane MCP

    Connect and sign in to Tetherplane first; the skill won't set it up. It works only on Windows PCs their owner has authorized. Browser tasks need a browser provider and tab sharing.

    Permissions

    Read Files
    Write Files
    Browser
    Terminal / Shell
    Listed11 days ago
    Updated10 days ago

    What's inside

    Frequently Asked Questions

    This skill acts as a specialized bridge between an AI agent and the Tetherplane MCP, replacing risky, open-ended shell commands with structured, scoped controls for files, processes, and browser automation on Windows.

    You must have an active, authenticated Tetherplane MCP installation on your target Windows host; the skill uses the compact interface to communicate across the 'device', 'files', and 'process' namespaces.

    Not by default. This skill does not grant administrative privileges or unrestricted filesystem access. It operates through your separately installed Tetherplane MCP agent, where device-local security policies, configured permissions, and operating-system privileges determine what actions are permitted. Administrative operations require appropriate authorization and existing system privileges. Filesystem tools enforce configured directory restrictions, while process execution has separate permissions and security considerations. The skill follows a read-first, least-privilege approach: inspect available permissions before acting, request approval where required, and never attempt to bypass denied operations. Installing the skill alone does not authorize access to your computer or its files.

    You receive the pre-configured skill definition designed for MCP-compatible agents, including the logic for device diagnostics, batch execution, and verified file operations.

    Unlike generic scripts, this skill uses specific tool calls to poll handles and verify exit codes, ensuring the agent only reports a task as successful once it has received byte-by-byte confirmation.

    No. It provides instructions to a compatible AI agent. Tetherplane itself must be separately installed, connected, and authorized.

    It follows the SKILL.md format. The agent must support the skill instructions and have access to an authenticated Tetherplane MCP connection. Compatibility depends on the client's integration.

    No. It can act only through devices and operations already authorized by their owners and local policies.

    Only when the browser provider is available and the appropriate permissions are granted. Human-owned tabs are protected by default.

    Yes, when those operations are supported and explicitly permitted by the connected device's policy.

    The skill instructs the agent to inspect connection and operation state before retrying. It must not blindly repeat actions that may already have completed.

    This listing is for a developer-beta skill. A successful demonstration does not establish production readiness or guarantee successful onboarding on every device.

    New in LLM Ops

    Free