Skip to content
sentrasec

AI coding agents

Security context where code is written.

An agent generating code has no view of your security posture unless you give it one. Sentrasec exposes first-class MCP tools so security becomes another capability the agent can use.

Why this matters now

A growing share of production code is written with an agent in the loop. That agent knows the codebase in front of it, but nothing about which dependencies carry known vulnerabilities, which patterns your rules reject, or what the last scan found.

Without that, review is the first point where a security problem surfaces: after the code exists and after the decisions are baked in.

Connected over MCP, the agent can scan what it just wrote, ask why a finding matters, and propose the fix before the code is ever committed.

How it connects

The agent talks to Sentrasec over MCP. Sentrasec answers from the same engine and graph that back every other surface.

  1. IDEagent
  2. MCPstdio
  3. Sentrasecengine
  4. Knowledge Graphcontext
  5. Responseinline
The editor calls a tool over MCP, Sentrasec runs the scan or traverses the graph, and the answer returns inline. Nothing about the agent's workflow changes: it simply gains a capability.

Available tools

Standard MCP tool schemas: your client discovers them after the server starts, and the agent picks the right one from a plain-language request.

ToolWhat it does
security.dast_scanRun a DAST scan against a URL. Returns deduplicated, enriched findings with business impact, technical detail and remediation.
security.list_findingsList findings for a scan, or across the whole workspace.
security.explainExplain a vulnerability by its knowledge-graph node id: description, CWE, OWASP mapping and remediation.
security.suggest_fixSuggest a code-level fix for a finding in a given language.
security.list_rulesList the detection rules available in the engine.
security.list_languagesList the languages the scanners support.
security.issue_certIssue a signed pentest certificate covering a set of scans.

Supported clients

Any MCP-capable client works. These are the ones with documented setup guides.

Claude Code

Add the server once; tools are discovered automatically in every session.

Cursor

Configure the MCP server in workspace settings and the agent gains the tools inline.

GitHub Copilot

Connect through the agent MCP configuration to bring findings into chat.

VS Code

Native MCP support wires the server into the editor agent.

Continue

Register the server in the assistant configuration alongside your models.

Claude Code
$ claude mcp add sentrasec -- sentrasec mcp

The server runs over stdio against your authenticated workspace: no network port is opened, and nothing is exposed beyond the machine the agent runs on.

Security as a capability, not a gate

Full MCP reference, tool schemas and per-client setup.

MCP documentation