Skip to content

Security That Lives Inside Your Tools

Integrations

Trent isn’t another dashboard to check. It plugs into your IDE, your AI app builder, and your agent runtime over MCP, then scans, prioritizes, and fixes from inside the workflow you already use. It connects to the tools your team already runs, and reads what they hold as evidence. Don’t have code yet? Point it at a URL or an idea.

What it adds

Judgment, not another console.

Not another dashboard or scanner: an engineer that works alongside the tools you already run.

Trent judges each finding against how your system actually runs, and tells you which few an attacker could use, with the evidence.

    • Code

      Applications and services

    • Infrastructure

      IaC, configs and scripts

    • Agents

      Skills, tools and LLMs 

    • Automations

      CI and GitHub Actions

    • Connected apps

      Your security tools and tickets 

    • Docs and ideas

      Designs, specs, requirements

  1. Context layer Continuously unified

  2. Your team AI Security Engineer Your coding agents

Your stack

Works With & Within Your Stack

Trent connects to any coding agent over remote MCP, installs as a plugin in Claude Code and Codex (built to the open Agent Plugins format, so other conformant clients can load it), runs as a skill inside OpenClaw, and takes your repo, live URL, or design docs directly.

Sources

Point Trent at what you're building: a repo, a live URL, or the design docs and ideas behind it. No code yet? Start with an idea.

  • GitHubConnect a repository.
  • Web URLAssess any live app, no code access required.
  • Code uploadBring your code directly.
  • Design docs & specsProduct specs, compliance requirements.
  • Agent filesAgents defined in markdown files.
  • IdeasNo code yet? Ask Trent’s security advisor before the first line is written.

Security connectors

Native connectors for Semgrep, Snyk and Sonar. Anything else comes in as SARIF. Trent shows those findings as evidence on every finding. Read-only, always.

More native connectors on the way. Tell us which one you need

  • SemgrepNative connector. One API token, findings stay current.
  • SnykNative connector. One API token, findings stay current.
  • SonarNative connector. One API token, findings stay current.
  • Any other tool via SARIFUpload SARIF from any scanner, pentest or cloud tool. Snyk and Semgrep JSON exports work too.

Team connectors

Trent reads the decisions and docs your team already wrote, and sends findings back as tickets and updates.

  • JiraTasks filed as tickets
  • LinearPull an issue in by its identifier. Decisions and accepted risks become context.
  • SlackKnow what changed after every push, get answers in the thread, and apply decisions with a button.
  • NotionDesign docs, threat models, pentest reports. Search, preview, add.

Coding agents

Bring Trent into the editor or agent where you write code. Plus any MCP-compatible client.

These are featured clients, not the full list. Trent supports MCP-compatible clients; if your tool speaks MCP, it can connect to Trent.

See the install steps

MCP Integration

Security Inside Your IDE and AI Tools

Trent’s MCP server runs inside Claude Code, Open AI’s Codex, Cursor, and most MCP-capable editors. Review the security of an idea even before a line of code exists, scan your repo, and apply prioritized fixes without leaving your coding environment.

  • Claude Code and Codex: install the Trent plugin in two commands, sign in with your browser, done. No API key, nothing to download. The plugin is built to the open Agent Plugins format, so any client that supports it, including Cursor and GitHub Copilot, can install it the same way.
  • Any MCP client, including Cursor, Windsurf, Gemini CLI, GitHub Copilot, Claude Desktop and Cowork: add Trent’s hosted server, then sign in with your browser.
  • Security review on ideas and design docs, before the code is written
  • Scan, prioritize, and mitigate from inside your editor
  • For OpenClaw, findings render right inside OpenClaw, where you already work
Install in Claude Code
claude plugin marketplace add trnt-ai/trent-agent-plugin
/plugin install trent@trent
claude mcp login trent-us
Install in Codex CLI
codex plugin marketplace add trnt-ai/trent-agent-plugin
codex plugin add trent@trent
codex mcp login trent-us

Sign in to trent-eu instead if your company’s data is held in Europe. The plugin is open source under MIT at github.com/trnt-ai/trent-agent-plugin.

Where fixes land

  1. A pull request

    Trent hands your coding agent the exact instructions, the agent prepares the pull request, and Trent reviews it for security and verifies the fix once it is merged.

  2. Your coding agent

    Start a remediation loop in your coding agent: it pulls Trent's tasks over MCP and opens pull requests for your review.

  3. A ticket

    Trent raises a ticket in Linear or Jira with exact remediation instructions.

Your terms

Runs wherever your security and compliance need it. Your code and context never leave your boundary.

Your boundary

Where it runs
  • Public cloud
  • Your VPC
  • On-prem
Which models
  • Frontier
  • Open-source
  • Private
Trust
  • SOC 2 certified

Agentic AI Security Platform Leader 2026 Cybersecurity Stars Awards

Next step

Secure what you build, wherever you build it.

Connect Trent to your IDE, your app builder, or your agent runtime, or just point it at a URL. Either way, security runs from the first idea to every deploy.

For AI agents

Using Claude Code, Codex, Cursor, or another coding agent? Point it at Trent’s machine-readable product and integration context.

Get secure.