Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    mcp prompt injection probe

    1

    Dynamic red-teaming skill for testing prompt injection vulnerabilities in live MCP server and agent setups.

    $18

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    mcp prompt injection probe

    mcp prompt injection probe

    Example session with this skill installed

    Run the starter battery against my staging environment. I have written authorization for the 'internal-crm' agent and 'git-mcp' server. Class B testing is enabled via the mock database.

    • Read your context and instructions
    • Compiled the mcp prompt injection

    Probe complete.
    Result: 6/8 Blocked, 1/8 Partial, 1/8 Inconclusive.
    Class B (Response Injection): Partial. The agent stated it would ignore the 'system override' instruction but still executed the 'delete_record' tool call. Evidence attached in transcript.log.

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

    About this skill

    The problem

    Static analysis cannot predict how an agent behaves when a live tool returns attacker-controlled content. Relying on documentation or tool manifests leaves your system exposed to runtime prompt injection that only surfaces during execution.

    What it does

    • Executes real-world adversarial payloads against live agent and MCP server configurations.
    • Tests four specific attack vectors: poisoned tool descriptions, injected tool responses, cross-server tool shadowing, and authority manipulation.
    • Captures full tool-call transcripts to identify Partial triggers where an agent claims to resist but still executes the payload.
    • Validates client-side runtime defenses, output sanitizers, and human-in-the-loop guardrails under actual attack conditions.

    Why this beats prompting it yourself

    This skill follows a structured scoring rubric and a battery of mutated payloads, including paraphrasing and encoding variants, that manual testing often misses. It systematically tests the response-injection class responsible for major CVEs in the MCP ecosystem which are difficult to rig by hand.

    Use cases

    • Red-teaming a new MCP server integration in a staging environment before deployment.
    • Verifying that a security patch actually closes a prompt injection hole post-incident.
    • Benchmarking an agent's resistance to cross-server shadowing when multiple tools are co-installed.
    • Generating runtime evidence for security audits that static code reviews cannot provide.

    Known limitations

    Requires explicit written authorization and a non-production environment. It is a dynamic probe, not a static pre-install scanner or a source code auditor.

    Requires explicit written authorization and a non-production environment. Not for pre-install vetting or server source code auditing.

    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 2 months ago

    Requires

    Standing Up A Live Target. Reviewing An MCP

    Permissions

    No special permissions declared or detected

    Listed2 months ago

    What's inside

    Frequently Asked Questions

    This skill acts as a dynamic red-teaming tool that executes adversarial payloads against live MCP setups to identify runtime vulnerabilities like tool-shadowing and poisoned responses that static code analysis often misses.

    It is designed for any agent architecture that utilizes the Model Context Protocol (MCP); it tests the interaction between the agent, the MCP server, and the tool outputs regardless of the specific underlying LLM.

    For safety and accuracy, this skill must only be used in isolated staging or development environments with explicit authorization, as it simulates real-world attacks to test system resilience.

    You receive the full probe skill package, including a battery of mutated adversarial payloads, the structured scoring rubric, and the transcript capture logic for audit reporting.

    The skill includes a range of sophisticated attack vectors, such as encoded variants and paraphrased injections, specifically designed to bypass basic filters and detect "partial triggers" where an agent might appear to resist but still fails.

    New in Agents & Orchestration