Hunt Control Plane — Agentic OS

Settings

configure the beast

Safety

Auto-Approve Plans Live Target Execution Active-Safe Testing Sandbox Enabled

Model & Budget

Router Mode Hunt Loop Fallback to Mock Max Budget (USD): $2.00 Approval TTL (min): 60

Model Providers

API Keys & Endpoints

Set provider API keys and base URLs

Discord Alerts

Webhook URL set via DISCORD_WEBHOOK_URL env var. Test it fires on critical findings.

Model Routing %

rep1 §5.2 — percentage split across providers per tier

Flash Tier (fast/cheap)

Pro Tier (deep reasoning)

Role Thinking

Reasoning mode override per agent role

AI Routing (auto-assign based on task)

Runtime internal

MCP tools 0

Engagements 0

Safety gated

Metrics

tokens · cost · scope · tools · findings

Loading metrics…

Agent Cockpit

0 jobs

Focused Agent

none selected

Agent Prompt Tuning

0 roles

Edit any agent's system prompt. Changes apply to the next agent spawn (no redeploy needed).

Loading prompts…

Agent Work Trace

0 steps

Checkpoints

0 saved

Model Invocations

routing ledger

Memory Records

three-tier ledger

Sandbox Runs

bubblewrap gate

Live Agent Timeline

0 events

MCP Calls

policy gated

HackerOne Programs

Loading programs…

Findings & Reports

0 findings

Loading reports…

Engagement Setup

scope first

Approval Queue

human gates

MCP Servers

trust registry

Autonomous Bug Bounty Operations

Engagement Graph

Statistics

Settings

Safety

Model & Budget

Model Providers

API Keys & Endpoints

Discord Alerts

Model Routing %

Role Thinking

Metrics

Agent Cockpit

Focused Agent

Agent Prompt Tuning

Agent Work Trace

Checkpoints

Model Invocations

Memory Records

Sandbox Runs

Live Agent Timeline

MCP Calls

HackerOne Programs

Findings & Reports

Engagement Setup

Approval Queue

MCP Servers

Autonomous Bug Bounty Operations

Safety

Model & Budget

Model Providers

API Keys & Endpoints

Discord Alerts

Model Routing %

Role Thinking

Chats & Thinking

Prompt this agent (operator directive — runs via SDK+CCR)