Conviso MCP Server

Introduction

The Conviso MCP Server is a connector that exposes Conviso Platform data and actions to an LLM through the Model Context Protocol (MCP). MCP lets external services register named capabilities (tools) so an MCP-compatible client — Claude Desktop, Claude Code CLI, Cursor, or any other — can ask the model to fetch live, authoritative security context and perform operations, instead of relying on cached knowledge.

The server ships 44 tools in two families:

• Read tools (33) — list and inspect companies, projects, assets, vulnerabilities, tickets, requirements and their project activities, applications, scan histories, SBOM / supply-chain components, AI-Pentest artifacts and executions, threat-model artifacts, plus security metrics, object-to-company lookup and deep links.
• Write tools (11) — a generic, allowlisted mutation engine (list_mutations → describe_mutation → execute_mutation) plus curated shortcuts for the most common writes (change issue status, create vulnerabilities / projects / assets / tickets, run DAST, trigger an AI-Pentest).

Everything the server can do is bounded by the Conviso Platform API Key you provide — the data returned and the operations allowed match that key's permissions. Write operations (except ticket creation) also honor the company's MCP write policy in Conviso Platform.

Quick install

Pick your client below and add the server, supplying your API key as CONVISO_API_KEY. Most examples run the published Node package with npx — no clone or build required. All setups below expose the same read and write tools.

Read and write. The server acts only within the permissions of the API key you provide.

Claude Code

Register the server once and it's available in every project:

claude mcp add -s user -e CONVISO_API_KEY=<your_api_key> conviso-mcp -- npx -y @convisoappsec/mcp

Check it with claude mcp list (or run /mcp inside a session). Swap -s user for -s project to scope it to the current repo (.mcp.json).

Claude Desktop

Easiest: open Settings → Extensions → Explore extensions, search Conviso MCP Server, click Install, then Configure and paste your API key. Full walkthrough: Marketplace steps.

Manual: edit claude_desktop_config.json (Settings → Developer → Edit Config):

{
  "mcpServers": {
    "conviso-mcp": {
      "command": "npx",
      "args": ["-y", "@convisoappsec/mcp"],
      "env": { "CONVISO_API_KEY": "your_api_key_here" }
    }
  }
}

Cursor

Add to .cursor/mcp.json (current project) or ~/.cursor/mcp.json (all projects):

{
  "mcpServers": {
    "conviso-mcp": {
      "command": "npx",
      "args": ["-y", "@convisoappsec/mcp"],
      "env": { "CONVISO_API_KEY": "your_api_key_here" }
    }
  }
}

Then enable conviso-mcp under Settings → MCP.

VS Code

Add to .vscode/mcp.json — note the servers key and the type field:

{
  "servers": {
    "conviso-mcp": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@convisoappsec/mcp"],
      "env": { "CONVISO_API_KEY": "your_api_key_here" }
    }
  }
}

Or add it straight from the terminal:

code --add-mcp '{"name":"conviso-mcp","command":"npx","args":["-y","@convisoappsec/mcp"],"env":{"CONVISO_API_KEY":"your_api_key_here"}}'

Codex

Add to ~/.codex/config.toml:

[mcp_servers.conviso-mcp]
command = "npx"
args = ["-y", "@convisoappsec/mcp"]
env = { CONVISO_API_KEY = "your_api_key_here" }

Gemini

Add to ~/.gemini/settings.json (all projects) or .gemini/settings.json (current project):

{
  "mcpServers": {
    "conviso-mcp": {
      "command": "npx",
      "args": ["-y", "@convisoappsec/mcp"],
      "env": { "CONVISO_API_KEY": "your_api_key_here" }
    }
  }
}

Or from the terminal:

gemini mcp add -e CONVISO_API_KEY=<your_api_key> conviso-mcp npx -y @convisoappsec/mcp

Works in Gemini CLI and Gemini Code Assist (agent mode). Verify with /mcp list. For HTTP transport, run the server with PORT=3000 and use "httpUrl": "http://localhost:3000" instead of command.

Docker

Run the Node image (build it first: docker build -t conviso-mcp-node-image node/):

{
  "mcpServers": {
    "conviso-mcp": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm", "--init",
        "-e", "CONVISO_API_KEY=your_api_key_here",
        "conviso-mcp-node-image"
      ]
    }
  }
}

Restart or reload your client after adding the entry. The tabs use the stdio transport; for Claude Cowork/Projects or remote hosting, use HTTP (Connector) mode below.

---

What you can do

Domain	Read	Write
Companies	List, inspect (plan, integrations, branding)	—
Vulnerabilities / Issues	List (rich filters), full technical detail, severity overview, per-asset and per-project views	Change status, create/update (source-code, web, network), reassign, mark analyzed, bulk status, bulk delete
Projects	List (filters), inspect detailed planning and allocation data, list project types and statuses	Create (with requirements and assets), update, change status, bulk status, bulk delete
Assets	List (filters), inspect	Create, update, run Conviso DAST
Tickets	List, inspect	Create
Requirements / Checklists	List by scope, inspect, list per project, list instantiated project activities	Create/update, attach to projects
Applications	List, inspect (with linked assets)	Create, update, add/remove assets
Scans	Execution history, coverage counts	(read-only)
Supply chain / SBOM	List components (license, versions, issues by severity)	(read-only)
AI-Pentest	List artifacts, inspect scope & executions, execution results	Create/update artifact, trigger execution, cancel, retest
Threat Modeling	List artifacts, inspect versions	Create artifact/version, update, generate requirements
Utilities, metrics & links	Resolve a company from an existing object, MTTR over time, risk-score history, deep links, today's date	—

Tools reference

Read tools

Domain	Tool	Description
General	get_companies	List companies accessible with the API key (search = name contains, label_eq = exact match).
General	get_company_info	Company detail: plan, integrations, branding metadata.
General	get_company_id_from_object	Resolve the owning company from an issue, asset, project, AI-Pentest artifact/execution, or threat-model artifact ID before a write.
Vulnerabilities	get_issues	List vulnerabilities with rich filtering and sorting; use asset_id or project_id to scope the results.
Vulnerabilities	get_issue	Full technical detail for one issue; optional vulnerable code snippet and raw HTTP request/response.
Vulnerabilities	get_top_vulnerabilities	Counts grouped by severity (risk overview), with optional filters.
Projects	get_projects	List security projects for a company, with filters and sorting.
Projects	get_project	Detailed project data: planning dates, effort, goals, scope, assigned users and teams, allocated hours, assets, requirement progress and project type.
Projects	get_project_types	List project types and find the type_id required by create_project; supports label search.
Projects	get_project_statuses	List valid project statuses (id, label, isInitial).
Assets	get_assets	List assets for a company, with rich filtering and sorting.
Assets	get_asset	Detail for a specific asset.
Tickets	get_tickets	List tickets (paginated) with search and TicketSearch params.
Tickets	get_ticket	One ticket: status, priority, impact, assignee.
Requirements	get_requirements	List requirements/checklists for a scope (company).
Requirements	get_requirement	One requirement/checklist by id.
Requirements	get_project_requirements	Requirements attached to a specific project.
Requirements	get_project_requirement_activities	Instantiated activities for a project requirement, including status, assignees and history count.
Applications	get_applications	List applications (name, url, riskScore, assetsCount).
Applications	get_application	One application, including its linked assets.
Scans	get_scan_histories	Scan executions (status, integration, duration, vuln counts).
Scans	get_asset_scans_count	Scan-coverage counts (assets with / without scans).
Supply chain	get_sbom_components	SBOM components: version, technology, package manager, license, issues by severity.
AI-Pentest	get_pentest_artifacts	List AI-Pentest artifacts (label, type, scheduling, latest execution).
AI-Pentest	get_pentest_artifact	One artifact, including scope and its executions.
AI-Pentest	get_pentest_execution	Execution result: status, vuln count, severity breakdown, retest progress.
Threat Modeling	get_threat_model_artifacts	List threat-model artifacts (label, scope, latest version).
Threat Modeling	get_threat_model_artifact	One artifact, including its versions (diagrams, notes, scope).
Metrics	get_mttr_over_time	MTTR aggregated over a date range; severity/status/asset filters.
Metrics	get_overall_risk_score_history	Historical risk scores for trend analysis.
Utilities	create_project_url	Deep link to a project in the Platform.
Utilities	create_issue_url	Deep link to a specific issue.
Utilities	get_today_date	Current date — useful to compute relative ranges ("last 30 days").

Write tools

Group	Tool	Description
Engine	list_mutations	Discover the permitted write operations (name, description, category, destructive flag).
Engine	describe_mutation	Full input schema for one mutation (fields, required, enums, nested inputs) + default return fields.
Engine	execute_mutation	Run any allowlisted mutation by name with a generic input object.
Shortcut	change_issue_status	Change an issue/vulnerability status.
Shortcut	create_source_code_vulnerability	Create a manual source-code (SAST-style) vulnerability on an asset.
Shortcut	create_project	Create a project, optionally associating requirements and assets. Use get_project_types to find the required type_id.
Shortcut	create_asset	Create an asset.
Shortcut	create_ticket	Open a ticket.
Shortcut	run_dast	Start a Conviso DAST scan on an asset.
Shortcut	trigger_pentest	Trigger an AI-Pentest execution from an existing artifact.
Shortcut	create_pentest_artifact	Create an AI-Pentest artifact (the scope/config a pentest runs against).

Prerequisites

• Conviso Platform API Key — create it under Profile > API Keys.
• An MCP-compatible client: Claude Desktop, Claude Code CLI, Cursor, or any client supporting stdio/HTTP MCP servers.
• Node.js 20.10+ or Docker.

Generate the key in the Conviso Platform under Profile > API Keys and copy it — you'll pass it to the server as the CONVISO_API_KEY value in the steps below.

Security recommendation

Create a dedicated API key for the MCP server and set an expiration date. Grant it only the permissions the workflow needs, and never commit it to version control.

Tips for better prompts

• Provide the company ID once you know it — it unlocks most tools. If you don't have it, start with "List my companies".
• For writes with only an object ID, let the model call get_company_id_from_object first instead of guessing the company ID.
• Ask for links at the end of any vulnerability or project query — the model calls create_issue_url / create_project_url as a follow-up.
• Combine context — the AI keeps conversation context, so "now get the details for the first one" works after a list.
• Request formats — ask for tables, bullet lists, or raw JSON ("format as a markdown table", "give me just the IDs").

---

Security and privacy

• The server operates strictly within the permissions of the provided API Key — for both reads and writes. Except for ticket creation, writes also require the company's MCP write policy to be enabled.
• Writes are allowlisted. Only the operations listed above are reachable; destructive ones (delete / bulk / cancel / remove) are flagged so the client prompts for confirmation.
• get_issue with return_vulnerable_data=true may return exploit code, raw HTTP requests/responses, or secrets — use with care.
• Create a dedicated key with an expiration date; do not reuse a personal API key, and scope it to the minimum permissions needed.
• Keep the API key out of version control — use environment variables or a secrets manager.

Privacy Policy

This connector communicates only with the Conviso Platform API (https://app.convisoappsec.com) using the API key you provide. It does not collect, store, or share your data with any third party — requests and responses stay between your MCP client and the Conviso Platform. Error logs go to stderr only.

Full privacy policy: iubenda.com/privacy-policy/55589285

Open source

The Conviso MCP Server is open source at github.com/convisoappsec/conviso-mcp. Contributions are welcome — bug reports, new tools, and improvements. See the repository's CONTRIBUTING.md to get started.

See also

• Conviso Skills — reusable operational playbooks for bulk actions (vulnerability triage, owner assignment, asset risk normalization) via conviso-cli, with preview-first safety controls.

Support

If you need help, contact Conviso support: support@convisoappsec.com.