codeloop_interact

// WHEN AI AGENTS USE THIS TOOL

AI agents invoke codeloop_interact to trigger processes or run actions in Codeloop. Execute operations can have side effects beyond the immediate call -- triggering builds, sending notifications, or starting workflows. Rate limits and argument validation are essential to prevent runaway execution.

// WHY ENFORCE A POLICY ON CODELOOP_INTERACT

codeloop_interact can trigger processes with real-world consequences. An uncontrolled agent might start dozens of builds, send mass notifications, or kick off expensive compute jobs. Intercept enforces rate limits and validates arguments to keep execution within safe bounds.

// RECOMMENDED POLICY

Execute tools trigger processes. Rate-limit and validate arguments to prevent unintended side effects.

codeloop.yaml

tools:
  codeloop_interact:
    rules:
      - action: allow
        rate_limit:
          max: 10
          window: 60
        validate:
          required_args: true

See the full Codeloop policy for all 29 tools.

// DETAILS

Tool Name codeloop_interact

Category Execute

MCP Server Codeloop MCP Server

Risk Level High

// MORE CODELOOP TOOLS

Execute codeloop_design_compare Execute codeloop_diagnose Execute codeloop_record_interaction Execute codeloop_run_history Execute codeloop_start_recording Execute codeloop_stop_recording Execute codeloop_visual_review Write codeloop_generate_dev_report

View all 29 tools →

// WHAT CAN GO WRONG

Agents calling execute-class tools like codeloop_interact have been implicated in these attack patterns. Read the full case and prevention policy for each:

Browse the full MCP Attack Database →

// OTHER EXECUTE TOOLS ACROSS MCP SERVERS

Other tools in the Execute risk category across the catalogue. The same policy patterns (rate-limit, validate) apply to each.

Atv build_stake_tx Atv build_unstake_tx Acdp start_local Pentagonal pentagonal_audit Pentagonal pentagonal_compile ACR — Agent Composition Records orient_me

// IN CONTEXT OF HIGH RISK

codeloop_interact is one of the high-risk operations in Codeloop. For the full severity-focused view — only the high-risk tools with their recommended policies — see the breakdown for this server, or browse all high-risk tools across every MCP server.

// RELATED READING

// FAQ

What does the codeloop_interact tool do? +

[CodeLoop] You MUST call codeloop_verify after every code change. If .codeloop/config.json is missing, call codeloop_init_project FIRST. Perform UI interactions on the running app during a recording session. Use this instead of raw osascript/PowerShell/xdotool commands. Supports desktop (macOS/Windows/Linux), browser (Playwright), Android emulator (adb), and iOS Simulator (simctl). IMPORTANT: Call this tool BETWEEN codeloop_start_recording and codeloop_stop_recording to actively interact with EVERY element in the app. Do NOT let the recording sit idle. WINDOW FOCUS: This tool automatically brings the target app to the front before each desktop interaction. The app name is auto-detected from the active recording session, or you can pass app_name explicitly. This ensures interactions hit the app window, not the IDE. Core actions: click, double_click, right_click, hover, type, keystroke, hotkey, scroll, drag_drop, long_press, type_and_submit, type_and_tab, fill_form, select_option, toggle, navigate_url, navigate_back, wait, sequence. navigate_url works on ALL targets: browser (Playwright), desktop (opens Chrome/Safari via osascript/PowerShell/xdg-open), Android (adb deep link), iOS (simctl openurl). scroll works on desktop via CGEvent (macOS), user32 mouse_event (Windows), xdotool (Linux). Falls back to arrow key presses if CGEvent fails (permissions). Browser-specific: Uses Playwright selectors (CSS/text) when target_type is "browser". Mobile-specific: swipe, back_button, home_button, deep_link, grant_permission, rotate_device, biometric_auth, launch_app, clear_app_data, mock_location, simulate_network. Maestro: maestro_flow — generate and run a Maestro YAML flow from high-level steps. Windows: win_ui_inspect, win_ui_automate — PowerShell UI Automation for UWP/WinUI apps. MANDATORY for web apps: You MUST type into form fields, fill login/signup forms, test validation errors, and click submit buttons. Just navigating pages is NOT enough. Wait 1-2 seconds between interactions so video frames capture state changes.. It is categorised as a Execute tool in the Codeloop MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on codeloop_interact? +

Add a rule in your Intercept YAML policy under the tools section for codeloop_interact. You can allow, deny, rate-limit, or validate arguments. Then run Intercept as a proxy in front of the Codeloop MCP server.

What risk level is codeloop_interact? +

codeloop_interact is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.

Can I rate-limit codeloop_interact? +

Yes. Add a rate_limit block to the codeloop_interact rule in your Intercept policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.

How do I block codeloop_interact completely? +

Set action: deny in the Intercept policy for codeloop_interact. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.

What MCP server provides codeloop_interact? +

codeloop_interact is provided by the Codeloop MCP server (codeloop-mcp-server). Intercept sits as a proxy in front of this server to enforce policies before tool calls reach the server.

Let agents act without letting them run wild.