Premium AI trading
command center.
A disciplined cockpit for PaperClip agents, product operators, human approvals, and safe execution gates. The system can move fast, but it respects read-only, draft-first, and approval-first boundaries.
Today's operating picture
The cockpit is designed to make the next best action obvious without loosening the safety model.
Owner command memo
Daily Owner Brief is the operating centerpiece: what changed, what needs approval, system health, and Top 3 priorities.
- Priority 1Review the Human-in-the-loop desk for anything that needs Rey's approval.
- Priority 2Use Bull HQ Director to synthesize agent outputs before committing to public or product actions.
- Priority 3Route trading-adjacent content through Risk Sentinel before any external-facing draft is used.
Safety posture
Agent Command Router
Route Telegram/iMessage-style commands into safe staged agent actions.
Hermes: ask Bull HQ Director what needs my approval today. Hermes: have Risk Sentinel review the latest draft. Hermes: check Trading Bull AI and Dendrix health read-only.
Commandable operators
Each card prepares a safe terminal command. High-risk actions remain gated by explicit approval; this dashboard does not expand permissions.
Approval queue
Review pending decisions before dispatch. Paper execution bridge dispatch still requires an approved item and preserves the audit trail.
| Approval | Decision | Status | Created | Reason | Actions |
|---|
Connected operating surfaces
A read-only system map for PaperClip, Hermes, the VPS, and product operators.
Trading Bull AI and Dendrix dashboards
Product-specific operating cards for app health, deployment posture, and safety boundaries.
Operating loops and output discipline
Each operator has a consistent daily loop, artifact folder, output shape, and escalation boundary.
Telegram and iMessage cadence
Morning brief, conditional midday approvals, evening status, and urgent red-health alerts.
Owner workflow, repo intelligence, product health, and social queue
Polish v2 turns the cockpit into a daily decision surface: Detail Pages, approval timeline, agent activity feed, system health timeline, command palette, and mobile-friendly view.
Owner workflow
Approval execution allowlist
No live trading, broker automation, public deploys, or customer messages. Approved actions are strict allowlist artifacts first.
PaperClip, Trading Bull AI, and Dendrix
Shows latest git status, recent commits, changed files, failing tests posture, open TODOs, open issues, deployment risk, architecture notes, and recommended next PR.
Trading Bull AI and Dendrix
App status, container status, latest logs, uptime, known risks, last deployment, last backup, current safety mode, and next recommended maintenance.
Draft-first content queue
daily post ideas, market education snippets, compliance-reviewed captions, and approval-required publish queue.
Leadership operating report
Summarizes system posture, agent runs, approvals, content drafted, product telemetry, repo changes, unresolved risks, and next-week priorities.
Read-only local change inspection
Inspects the dirty local Trading Bull AI checkout without committing, discarding, or deploying anything.
Approval-gated product release plans
Pre-deploy checklist, latest commit, test/build status, telemetry, required approvals, rollback, deploy command preview, and post-deploy verification. No deployment is executed here.
Draft product release notes
Draft-only notes covering what changed, user/admin notes, risk, testing evidence, rollback notes, and owner approval. No customer message or public changelog is sent.
Approve, deny, defer, ask for plan, or attach artifacts
A cockpit-level decision board that makes risk, affected systems, exact action, attachments, timeline, and next step visible before execution.
Source-of-truth context for agents
company doctrine, Velez checklist, risk rules, brand voice, approved disclaimers, product notes, deployment notes, FAQ, and content library.
Tailscale first, Cloudflare Access second, never public
The cockpit stays local-only now. This plan documents future private access without exposing PaperClip publicly.
Decision artifacts
Repeatable briefing products that turn agent work into owner-ready decisions.
Search artifacts
Search saved agent run artifacts and owner briefs without leaving the local cockpit.
Recent cockpit signals
High-level status signals only. Use profile sessions and repo artifacts for full audit trails.
Draft review queue
Draft status, channel target, compliance status, risk level, revision notes, scheduled date, and owner approval are visible here. No publishing happens from this cockpit.