MCP Tools/List Change Reviewer

    1

    Compare two local MCP tools/list snapshots and review added, removed or changed tool descriptors before approving an update.

    $7

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    MCP Tools/List Change Reviewer

    Example session with this skill installed

    Compare the included synthetic previous.json and current.json MCP tools/list snapshots with MCP Tool Approval Change Card. Show the advertised descriptor changes and a human review checklist. Do not call any listed tool or grant permission.

    • Read your context and instructions
    • Compiled the mcp tools list
    • Generated the data export

    Synthetic demonstration using the sample files included in the download. The local analyzer's checked human-readable sample report is reproduced in full below; this is not a live account check or a platform-runtime test.

    MCP tool approval change review

    This packet compares advertised descriptors only. It does not inspect runtime behavior or grant approval.

    Previous tools: 2; current tools: 2.

    Changes

    • ADDED: 1
    • CHANGED: 1
    • REMOVED: 1

    Human signoff steps

    • Inspect original descriptions and schemas for every changed tool
    • Confirm actual server implementation and deployment identity
    • Treat annotation hints as untrusted; confirm real side effects and host approval policy
    • Record a human decision outside this packet before continued use

    A stable descriptor does not prove stable implementation. Read the source snapshots locally; this output stores hashes instead of original descriptions.

    Evidence in this sample

    The complete sample change queue records lookup as CHANGED (description hash and openWorldHint), new_export as ADDED, and old_report as REMOVED. All three rows require human review. The included change-cards.json and change-queue.csv contain the changed-field names and hashes; they do not claim that a tool implementation or deployment changed. The evidence receipt binds the local inputs and outputs.

    Next actions

    Inspect the original descriptions and schemas, verify the actual server implementation and deployment identity, and make any approval decision through your own host policy. The annotation hints in a tools/list response are self-declared; this card never grants access.

    change-queue.csv

    CSV · data export

    Generated

    Example file from a real run - the skill writes it into your workspace.

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

    About this skill

    An MCP server update can change the tools an agent sees without making those changes obvious during approval. This local tool compares two supplied tools/list snapshots and creates human review cards for added, removed and changed advertised descriptors. It does not connect to the server, execute tools or certify runtime safety.

    Compare two user-supplied local tools/list JSON snapshots and create review cards for added, removed or changed advertised tool descriptors. The local CLI writes a review packet and hash receipt. The download contains readable Python source, a companion Skill, a local stdio MCP adapter, setup instructions and synthetic samples. Python 3.11 or newer is required; MCP use requires the separately installed optional mcp==2.2.0 package. It never connects to a live MCP server, calls listed tools, grants permissions or verifies runtime behavior. Tool annotations are self-declared and may be wrong; a review card is a handoff for an operator, not a security certification or automated approval.

    Compares advertised descriptors only. Stable text does not prove stable implementation. Self-declared annotations may be wrong; inspect actual deployment and use your host approval process. Requires local Python 3.11+ source installation.

    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 any payment.

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    Requires

    Python
    Approval Card MCP
    Tool Approval Card MCP

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    Reads only two buyer-selected local tools/list snapshots and writes a local review packet. Does not call tools, connect to servers, grant permissions, or verify runtime behavior.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    This skill automates the comparison of two MCP tool snapshots, generating a structured review card that highlights additions, removals, or changes in tool descriptions to simplify human auditing.

    The package includes the Python source code, a companion Skill, a local stdio MCP adapter, setup instructions, and synthetic samples to help you get started immediately.

    No, this tool is designed for offline analysis; it never connects to live MCP servers or executes any tools, serving strictly as a comparison and reporting layer for tool metadata.

    You will need Python 3.11 or newer; if you intend to use the MCP features, the optional 'mcp==2.2.0' package must be installed separately.

    The skill generates a handoff card for human review based on self-declared tool annotations, but it does not perform runtime security certifications or automated permission granting.

    Yes, it is designed to be environment-agnostic as long as you provide the local tools/list JSON snapshots for comparison, making it compatible with various AI agent workflows.

    New in Agents & Orchestration