More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    🛡️ GuardrailDoctor

    2

    Review Claude Code hooks, spot configuration gaps, and plan controlled checks with evidence you can inspect.

    $29

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    🛡️ GuardrailDoctor

    🛡️ GuardrailDoctor

    Example session with this skill installed

    Review the settings and hook source I selected without executing hooks. Separate static findings from runtime checks that have not been run.

    • Read your context and instructions
    • Compiled the guardraildoctor

    Illustrative report format, not a completed live test

    The report separates what can be read from configuration and source from what must be tested in an approved host session. This example does not claim that any production guardrail passed.

    Review areaEvidence to recordBoundary
    Hook configurationSelected file and cited settingDeclaration is not proof of execution
    Response handlingEvent-specific response fieldsStatic review only
    Sensitive-data handlingWhether source records raw inputsNever probe real credentials
    Runtime enforcementNOT_RUN until authorized testingNo pass inferred from unit tests

    Next steps

    Review cited findings, back up customizations, and approve any proposed changes separately. If runtime checks are needed, prepare an isolated project with synthetic fixtures, record the actual invocation and result, and retain unresolved checks as NOT_RUN.

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

    About this skill

    Review before you rely on a guardrail

    GuardrailDoctor guides your agent through a review of selected Claude Code settings and hook source. Separate static findings from runtime evidence, record what still needs checking, and plan controlled tests. It is not an autonomous penetration-test runner or a safety certification.

    Current package

    Package v1.1.0 is listed in Agensi revision v1.2. It includes static-review-first instructions, event-specific references, an illustrative fixture hook and local tests. Inspect the version before updating an older installation.

    Safe first use

    Read SKILL.md and ask your agent to review owner-selected settings and source without executing hooks. Runtime testing requires explicit approval, an isolated project and synthetic data. Do not use real credentials as test probes.

    What the update changes

    The v1.1.0 fixture helper does not execute submitted commands, grant unmatched actions or log raw tool inputs. It is a teaching and testing aid, not a production security boundary. Packaged unit tests do not establish live host enforcement.

    Installation

    Extract the complete guardrail-doctor folder. An older installation may be named guardraildoctor. Inspect existing installations and back up customizations; avoid two active copies. Do not automatically delete or overwrite customized installations. Host and model charges are separate.

    Agent-guided static review, not an autonomous penetration-test runner or safety certification. Runtime enforcement remains NOT_RUN until separately tested. Fixture tests do not prove a live host blocks attacks. Never probe real credentials or treat illustrative output as a completed audit.

    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

    Permissions

    Terminal / Shell
    Read Files

    File Scopes

    .claude/settings.json
    .claude/hooks/**/*
    **/.*rc
    **/*.env

    Guardrail Doctor needs terminal access to safely test hook configurations with benign test payloads (e.g., harmless echo commands). It never executes destructive or unauthorized commands. File read access allows it to audit existing security settings. The skill does not write to any files—all recommendations are provided as output for the user to review and apply manually, ensuring the user remains in control of their security configuration.

    Listed5 months ago
    Updated1 month ago

    What's inside

    Frequently Asked Questions

    GuardrailDoctor automates adversarial testing by attempting to bypass your security hooks using techniques like shell substitution and path traversal, identifying vulnerabilities that manual testing often misses.

    This skill is specifically optimized for Claude Code environments and supports the auditing of .claude/settings.json configurations as well as custom PreToolUse and PostToolUse security hooks.

    You will receive the GuardrailDoctor skill package, a suite of standardized penetration testing payloads, and a structured reporting template that provides actionable remediation code.

    No, GuardrailDoctor is designed to be plug-and-play; as long as you have Claude Code installed, the skill can begin auditing your existing Bash, Read, Write, and MCP tool configurations immediately.

    The skill generates a comprehensive audit report featuring a pass/fail table for 10+ attack vectors and provides specific, copy-pasteable snippets to patch any identified security gaps.

    New in Security