BlastRadius-Zero: Pre-Flight Safety & Guardrails for Agents

    1

    Pre-flight safety runtime that scores risk, blocks dangerous commands, and redacts secrets before execution.

    Free

    1 useSecurity scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    BlastRadius-Zero: Pre-Flight Safety & Guardrails for Agents

    Example session with this skill installed

    Please clean up our staging environment: delete all obsolete folders in the root directory, drop the staging postgres cache table, and wipe any temporary S3 assets.

    • Read your context and instructions
    • Compiled the blastradius-zero

    🛡️ BlastRadius-Zero Pre-Flight Safety Interception

    Action Intercepted: Destructive Command Execution
    Status: 🛑 BLOCKED (Action Refused)

    Pre-Flight Simulation Report
    {
      "dangerScore": 100,
      "severity": "CRITICAL",
      "category": "SHELL & CLOUD",
      "reasons": [
        "Recursive wildcard filesystem deletion attempted across root targets.",
        "Destructive database query 'DROP TABLE' violates Zero-Trust Rule ZT-003.",
        "Cloud storage mass deletion request detected on target 's3://staging-assets'."
      ],
      "destructive": true,
      "irreversible": true,
      "rollbackFeasible": false,
      "recommendedMitigations": [
        "Specify explicit build/cache directories (e.g., ./dist or ./cache) rather than root paths.",
        "Use TRUNCATE or soft-delete partitions rather than dropping schema tables.",
        "Generate an isolated dry-run backup prior to modifying S3 buckets."
      ]
    }
    

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

    About this skill

    What is BlastRadius-Zero?

    BlastRadius-Zero is the pre-flight zero-trust security runtime and context-optimization gateway for AI coding agents.

    When you give an agent access to bash, filesystems, SQL, or cloud environments, it possesses the power to execute catastrophic commands (rm -rf, DROP TABLE, aws s3 rb). Traditional regex guardrails fail because bash transforms commands before execution. BlastRadius intercepts calls pre-flight, simulates shell resolution, calculates blast radius danger scores (0–100), redacts credentials flowing in both directions, and stress-tests code diffs against 50 adversarial virtual personas in <50ms.

    Key Capabilities

    • Pre-Flight Blast Radius Simulation: Calculates risk severity (SAFE, ELEVATED, CRITICAL) before destructive commands run.
    • Zero-Trust Command Invariants: Blocks root/wildcard deletions, unreviewed piping (curl | bash), and destructive SQL queries.
    • Secret Redaction (DLP): Automatically masks AWS keys, GitHub PATs, JWTs, and passwords.
    • Step-Up Cryptographic Tokens: Prevents token replay and enforces explicit human-in-the-loop elevation for high-risk actions.
    • Tamper-Evident Audit Ledger: Records every tool call to a hash-chained local SQLite database.

    Supported Agents

    Works natively with Claude Code, Cursor, GitHub Copilot, Codex CLI, Windsurf, and Cline.

    Quickstart

    Install and run the underlying MCP server locally:

    {
      "mcpServers": {
        "blastradius": {
          "command": "npx",
          "args": ["-y", "blastradius-mcp"]
        }
      }
    }
    

    Requires Node.js >= 22.0.0 and the blastradius-mcp server configured in the agent runtime. Proxy scoring is simulation-based and relies on tool-provided command strings.

    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

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 29 days ago

    • Free with an Agensi account

    Requires

    Node.js
    Blastradius MCP

    Permissions

    Write Files
    Terminal / Shell
    Read Files
    Listed29 days ago

    What's inside

    Frequently Asked Questions

    BlastRadius MCP acts as a zero-trust proxy that intercepts commands from your AI agent to score their danger, redact sensitive PII/secrets, and block destructive actions like unauthorized file deletions or database drops.

    This skill is compatible with any AI agent or platform that supports the Model Context Protocol (MCP). It requires a Node.js environment (version 22 or higher) to host the BlastRadius server layer.

    Unlike system prompts, which are vulnerable to jailbreaking and prompt injection, this skill provides a programmatic execution layer that operates independently of the LLM's logic to enforce hard security boundaries.

    For high-risk operations, the skill triggers a 'Step-up Auth' flow that generates a single-use cryptographic token, requiring a manual human confirmation before the command can proceed.

    You will receive the BlastRadius MCP server package, configuration templates for zero-trust policies, and full documentation for integrating the security proxy into your existing MCP-enabled agent workflow.

    While it provides robust logic-based filtering and command scoring, it is intended to sit on top of your existing OS-level and database permissions, not replace them.

    New in Security