mcpguard

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+22 more

    mcpguard

    1

    Audit MCP servers for prompt injection, secret leaks, and filesystem violations using MCPGuard.

    $6

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    mcpguard

    Example session with this skill installed

    Run a security scan on the 'sqlite-explorer' MCP server tool using the default policy. The server command is 'npx @modelcontextprotocol/server-sqlite --db-path ./data.db'.

    • Read your context and instructions
    • Compiled the mcpguard

    I have performed a targeted security scan on the sqlite-explorer tool. The audit checked for prompt injection, secret leaks, and basic schema integrity using the MCPGuard CLI.

    MCPGuard Security Assessment

    MetricDetails
    Targetnpx @modelcontextprotocol/server-sqlite
    Tool Scopesqlite-explorer
    Transportstdio
    Exit Code1 (Failed Severity Gate)
    StatusBLOCKED

    Confirmed Findings

    • High: Prompt Injection Vulnerability
      The sqlite-explorer tool accepted a generated input that resulted in an unauthorized schema dump. The tool does not sufficiently sanitize inputs before passing them to the database engine.
    • Medium: Path Disclosure
      Error messages from the server revealed the absolute internal path of the database file ./data.db, violating filesystem boundary expectations.
    • Info: Secret Scan
      No hardcoded credentials or API keys were detected in the tool's initial handshake or output.

    Coverage Summary

    • Tests Executed: Fuzzing, Injection, Secret-Leak, Schema-Validation.
    • Tests Skipped: Network-Boundary (Not implemented), Side-Effect-Detection.

    Next steps

    • Review the sqlite-explorer implementation to add parameterized queries or input validation.
    • Run mcpguard test --command "..." --output json to get the full trace for the injection finding.
    • Update the local policy to explicitly restrict filesystem access to the data directory only.

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

    About this skill

    Unvetted Model Context Protocol (MCP) servers pose significant security risks, including secret leaks, prompt injection, and unauthorized filesystem access. This skill provides a specialized interface for auditing and testing MCP servers using the MCPGuard CLI to ensure tools meet safety requirements before deployment.

    What it does

    • Pre-flight security testing executes active audits against MCP server endpoints to identify vulnerabilities.
    • Vulnerability detection identifies secret leaks, prompt injection risks, and filesystem boundary violations.
    • Policy enforcement validates server configurations against custom failure thresholds and timeout rules.
    • Evidence-based reporting generates trust-gate reports in terminal, JSON, or SARIF formats for CI/CD integration.
    • Target discovery inspects transport layers, tool scopes, and policy files to map the attack surface.

    How it works

    1. Target Identification involves locating the server via command, URL, or local registry and inspecting its tool scope.
    2. Policy Configuration reviews path boundaries, secret patterns, and failure thresholds to define the test gate.
    3. Execution Control runs targeted scans on specific tools or full-server audits using the mcpguard test command.
    4. Result Interpretation analyzes exit codes and heuristic findings to categorize risks by severity.

    Frameworks & tools

    This skill requires python 3.12 or newer and the mcpguard CLI. It supports MCP servers using standard transport protocols and integrates with security tooling via SARIF output.

    Why this beats prompting it yourself

    Generic agents often overlook the side effects of active testing or fail to interpret specific MCP schema errors correctly. This skill follows a strict safety workflow that prevents accidental production scans and ensures findings are backed by exit codes and heuristic evidence rather than LLM hallucinations.

    Use cases

    • Security Auditing to verify a new third-party MCP tool before adding it to an agent's environment.
    • CI/CD Integration to block builds if an MCP server fails secret-leak or injection tests.
    • Configuration Review to inspect tool path boundaries and network policy placeholders.
    • Compliance Reporting to generate SARIF evidence for security teams during release cycles.

    Known limitations

    Does not provide process, filesystem, or network isolation. Network-boundary enforcement and side-effect detection are currently not implemented in the underlying tool.

    How to use

    Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill passes our trust checks below.

    Security scanned

    Passed 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 scanned

    Verified clean 1 day ago

    • Passed all security checks, Safe to install

    Requires

    Python

    Listed1 day ago

    What's inside

    Frequently Asked Questions