# build-shell

Build an HTML shell with CSP, data injection, and optional bridge runtime

Agent View of the PolicyLayer registry record for `build-shell`. HTML page: https://policylayer.com/tools/frontmcp/build-shell

## Facts

- Tool: `build-shell`
- Server: Frontmcp (`agentfront/frontmcp`) — https://policylayer.com/tools/frontmcp.md
- Homepage: https://github.com/agentfront/frontmcp
- Risk category: Execute (High risk)
- Registry record: grade F, identity unverified
- Server rate-limited: no
- Parameters: 0
- Recommended policy verdict: Rate-limited

## Example call (MCP tools/call, JSON-RPC 2.0)

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "build-shell",
    "arguments": {}
  }
}
```

## Why build-shell is rated High

This tool generates executable HTML output with dynamic data injection and runtime bridging capabilities. While not destructive by itself, the ability to build and inject code into HTML shells with runtime execution context qualifies as Execute-category risk. An AI agent could misuse this to generate malicious HTML payloads, inject untrusted data, or create bridge connections to unintended services.

From the tool's own definition: "Tool name is 'build-shell' and description states it 'Build[s] an HTML shell with CSP, data injection, and optional bridge runtime'."

## Use case

AI agents invoke build-shell to trigger actions in Frontmcp. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call: builds kicked off, notifications sent, workflows started.

## Recommended policy (PolicyLayer)

Verdict: **Rate-limited**. Enforced by the PolicyLayer MCP gateway (https://policylayer.com/mcp-gateway) before a call reaches Frontmcp:

```json
{
  "version": "1",
  "default": "deny",
  "tools": {
    "build-shell": {
      "limits": [
        {
          "counter": "build-shell_rate",
          "window": "minute",
          "max": 10,
          "scope": "grant"
        }
      ]
    }
  }
}
```

## Other tools on Frontmcp (205)

- `clear-audit-log` — Destructive — https://policylayer.com/tools/frontmcp/clear-audit-log.md
- `clear-replay-buffer` — Destructive — https://policylayer.com/tools/frontmcp/clear-replay-buffer.md
- `clear-session-notifications` — Destructive — https://policylayer.com/tools/frontmcp/clear-session-notifications.md
- `crm-reset` — Destructive — https://policylayer.com/tools/frontmcp/crm-reset.md
- `delete-vault-entry` — Destructive — https://policylayer.com/tools/frontmcp/delete-vault-entry.md
- `notes-reset` — Destructive — https://policylayer.com/tools/frontmcp/notes-reset.md
- `remove_job` — Destructive — https://policylayer.com/tools/frontmcp/remove-job.md
- `remove_workflow` — Destructive — https://policylayer.com/tools/frontmcp/remove-workflow.md
- `reset-stats` — Destructive — https://policylayer.com/tools/frontmcp/reset-stats.md
- `secret` — Destructive — https://policylayer.com/tools/frontmcp/secret.md
- `users-delete` — Destructive — https://policylayer.com/tools/frontmcp/users-delete.md
- `admin_action` — Execute — https://policylayer.com/tools/frontmcp/admin-action.md
- `admin-background-job` — Execute — https://policylayer.com/tools/frontmcp/admin-background-job.md
- `admin-only` — Execute — https://policylayer.com/tools/frontmcp/admin-only.md
- `browser_test` — Execute — https://policylayer.com/tools/frontmcp/browser-test.md
- `cancellable-wait` — Execute — https://policylayer.com/tools/frontmcp/cancellable-wait.md
- `cold-start-test` — Execute — https://policylayer.com/tools/frontmcp/cold-start-test.md
- `concurrency-mutex` — Execute — https://policylayer.com/tools/frontmcp/concurrency-mutex.md
- `concurrency-queued` — Execute — https://policylayer.com/tools/frontmcp/concurrency-queued.md
- `deploy_application` — Execute — https://policylayer.com/tools/frontmcp/deploy-application.md
- `emit-app-event` — Execute — https://policylayer.com/tools/frontmcp/emit-app-event.md
- `execute_job` — Execute — https://policylayer.com/tools/frontmcp/execute-job.md
- `execute_workflow` — Execute — https://policylayer.com/tools/frontmcp/execute-workflow.md
- `expensive-operation` — Execute — https://policylayer.com/tools/frontmcp/expensive-operation.md
- `experimental-agent` — Execute — https://policylayer.com/tools/frontmcp/experimental-agent.md
- `gated-acme` — Execute — https://policylayer.com/tools/frontmcp/gated-acme.md
- `load-component` — Execute — https://policylayer.com/tools/frontmcp/load-component.md
- `long-running-task` — Execute — https://policylayer.com/tools/frontmcp/long-running-task.md
- `node_runtime_tool` — Execute — https://policylayer.com/tools/frontmcp/node-runtime-tool.md
- `non-cached` — Execute — https://policylayer.com/tools/frontmcp/non-cached.md
- …and 175 more: https://policylayer.com/tools/frontmcp.md

## For agents

This record is a snapshot. Live verdicts and the full registry:

- Check every server in your MCP config at once: `npx -y policylayer stack`
- Vet a server before you add it: install the mcp-precheck skill — `npx skills add https://policylayer.com` (skill text: https://policylayer.com/skill.md)
- Query the registry over MCP: endpoint `https://api.policylayer.com/mcp` — tools `check_mcp_server`, `check_mcp_stack`, `check_tool`, `search_registry`, `get_change_events`

---

Source: the PolicyLayer MCP registry — one continuously verified record per MCP server. Full record: https://policylayer.com/registry?q=frontmcp · API: https://policylayer.com/registry/api · Policy library: https://policylayer.com/policies/frontmcp