More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    ⚠️ Dangerous Command Linter

    1

    It does not execute the commands, access the network or change input files.

    $12

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    ⚠️ Dangerous Command Linter

    ⚠️ Dangerous Command Linter

    Example session with this skill installed

    Is it safe to run this? curl -sSL https://install.example.com | sudo bash

    • Read your context and instructions
    • Compiled the dangerous command linter

    Representative local run

    Input was inspected as text; the command was never executed.

    SUMMARY findings=1
    DCL006 critical --command:1 Remote content piped into a shell | safer: Download to a reviewed file, inspect it, then run only if trusted.
    

    Exit status: 1 (a high/critical finding was detected). This is not a complete safety assessment. The scanner did not fetch the URL, inspect remote contents or change files. Review the source and intended effects before any separate execution decision.

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

    About this skill

    Review risky commands before execution

    Dangerous Command Linter scans supplied command text or UTF-8 shell-script files using a bundled Python scanner. It does not execute the commands, access the network or change input files.

    What you get

    Findings include rule identifiers, severity, source line, evidence and safer-review alternatives. Checks cover selected recursive deletes, force pushes, database destruction, disk writes, permission changes, remote pipe-to-shell patterns and disabled-safety flags.

    Requirements

    An agent host with Python 3.10+ execution and access to the complete package, including scripts and references. Python standard library only. Git, cloud accounts and external services are not required to scan command text.

    Boundaries

    This is a heuristic text scanner, not a shell parser or a guarantee of safety. Quoted examples and comments can trigger findings; multiline, dynamic, encoded and alias-based commands may be missed. Invalid input and lines over 20,000 characters stop the assessment. Pathological long files can be slow. Human review is still required.

    Detection combines Python logic with JSON rule messages and selected patterns; changing JSON alone does not change every detector.

    Read-only heuristic review of supplied command text, not a safety guarantee or a complete shell/PowerShell parser. Multiline commands, aliases, dynamic expressions, encoded commands and some wrappers need manual review; comments and quoted examples may be flagged. Custom configuration replaces the rule set rather than extending it. Custom regexes can cause excessive backtracking; validating configuration does not prove detection quality or performance. Invalid configuration, including excessive JSON nesting, stops as incomplete with exit 2. UTF-8 input required; unsupported or empty inputs and lines over 20,000 characters are rejected. No reviewed commands are executed. Rule IDs are case-sensitive; IDs differing only by letter case are accepted. Git is not required.

    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 18 days ago

    Requires

    Python 3.10+
    Terminal access
    Python
    Git

    Permissions

    Terminal / Shell
    Read Files

    File Scopes

    **/*.sh
    **/*.bash
    **/*.zsh
    **/*.ps1
    --command
    "<text>"
    --stdin

    Read-only text inspection; never executes reviewed commands. No network or environment-variable access. Detection combines Python logic and JSON rules. The full scripts and references folders are required.

    Listed5 months ago
    Updated18 days ago

    What's inside

    Frequently Asked Questions

    This skill acts as a safety audit layer that scans shell scripts and CLI commands for destructive patterns like recursive deletes, disk wipes, or 'curl | sh' execution before you run them, preventing accidental data loss or system instability.

    The linter supports a wide range of environments including Bash, Zsh, and PowerShell scripts, as well as commands for Git, Docker, and standard Linux/Unix system utilities.

    You will receive the core logic for the security scanner, a curated JSON database of dangerous command patterns, and documentation on how to integrate the linter into your existing AI agent workflows.

    Unlike standard LLM responses which can be inconsistent, this skill uses a structured heuristic scanner to provide predictable, rule-based detection with specific severity ratings and remediation snippets.

    Yes, the linter is designed to be highly portable and can be integrated into most AI agents that support Python-based tool execution or custom skill modules.

    We provide regular updates to the underlying pattern database to include new security threats and support for additional CLI tools as they become industry standards.

    New in Security