Concept Exploration — gblock-party

A persistent managed development environment that runs multiple Codex CLI sessions in parallel, controlled from browser or phone. Five integrated surfaces — Agent Board, Terminal, VS Code, Browser Preview, and File Explorer — provide a complete AI coding workstation with zero public inbound access.

This is the personal dogfood environment. Agent Board + Terminal are the priority surfaces. VS Code, Preview, and File Explorer are deferred to existing tools.

Running AI coding agents locally is constrained by:

• Resource limits: laptop CPU/RAM caps out at 3-5 parallel agent sessions.
• No mobile access: can't monitor, approve, or review agent work from phone.
• No orchestration UX: no unified view of what multiple agents are doing, their status, or their outputs.
• Session fragility: closing a laptop kills sessions; tmux helps but has no GUI.
• Security gaps: agents run with full local access, no guardrails on sensitive files.

The sole beneficiary is you — a solo developer who wants to run 5-12 parallel AI coding sessions across multiple repos, monitor them from Mac and iPhone, and maintain a locked-down security posture.

Human-in-the-loop multi-agent orchestration from any device.

• See all active agent sessions in one view with real-time status.
• Start, pause, resume, and kill sessions from browser or phone.
• Approve or reject risky actions (package installs, file reads, git pushes).
• Preview running apps alongside the agents building them.
• Review diffs before pushing.

Development Priority

Agent Board + Terminal first. These are the two most critical surfaces. VS Code, Browser Preview, and File Explorer are deferred — existing tools (code-server, direct Tailscale tunnel) cover those needs until the core orchestration UX is proven.

• Enterprise compliance or certifications
• Multi-user or team features
• Kubernetes or container orchestration
• Multi-cloud or autoscaling beyond CCX33/CCX43
• Native iOS app (PWA only)
• Model provider billing or API key resale
• Public SSH or publicly exposed web ports
• GitHub Actions CI/CD
• Full remote desktop experience

Confirmed Assumptions

• Hetzner CCX33 provides sufficient compute for 5-10 parallel agent sessions. [internal architecture decision, from prompt]
• Tailscale + Cloudflare Tunnel provides adequate dual-layer access. [resolved Q2]
• tmux is the right session management primitive. [from prompt]
• Agent Board + Terminal are the priority surfaces. [resolved Q1]

Unknowns Requiring Validation

Key Codex CLI capabilities relevant to the Agent Board:

• codex exec "prompt" — Run a prompt non-interactively (alias: codex e)
• --json — Print events to stdout as JSONL (state observation, output streaming)
• -o, --output-last-message FILE — Write agent's final message to a file
• --output-schema FILE — Constrain final response shape with JSON Schema
• -s, --sandbox MODE — Sandbox policy: read-only, workspace-write, danger-full-access
• --dangerously-bypass-approvals-and-sandbox — Skip prompts (for externally sandboxed envs)
• --ephemeral — Don't persist session files
• --color never — Disable color output (for piping)
• -m, --model MODEL — Select model
• -i, --image FILE — Attach image(s)
• -p PROFILE — Select config profile (NOT equivalent to Claude Code's -p)

Note: Codex's -p selects a configuration profile from config.toml, not print/prompt mode. The equivalent of Claude Code's -p is codex exec. Stdin is also supported — if no prompt argument is given, instructions are read from stdin.

Resolved sequencing: Business discovery first, then technical spike.

1. /pack install business-discovery — install research skills for commercial concept.
2. /icp — validate beneficiary hypotheses (starting with power users / solo founders).
3. Technical spike on codex exec --json — prove feasibility of programmatic control before any UI work.
4. /feature-interview on Agent Board + Terminal once CLI strategy is confirmed.

A browser/mobile SaaS control plane for AI coding agents on gblock-party's managed cloud infrastructure. Users connect their preferred front-end tool (T3 Code, Multica, Claude Code CLI) or use the native Agent Board.

All infrastructure is managed by gblock-party.

Builds on the personal workstation codebase via a shared monorepo.

Developers increasingly use AI coding agents but face these problems at scale:

• No remote orchestration: agents run in local terminals with no mobile or browser access.
• No multi-agent visibility: 5+ agents means 5+ terminal tabs, no unified status view.
• No safe approval workflow: permission prompts require physical presence at the machine.
• No session persistence: closing a connection loses context.
• Security gaps: agents have full machine access, no guardrails.
• Setup burden: configuring a secure remote dev environment requires significant sysadmin effort.

Primary ICP starting hypothesis: Power users / solo founders

To be validated by ICP analysis. Other segments for later investigation:

• Freelancers / agency developers: multiple client projects, session isolation, mobile access.
• Security-conscious developers: zero-inbound architecture, audit trails.
• Small teams: shared visibility without full machine exposure.

Secondary (managed IaaS tier)

• Non-sysadmin developers: turnkey devboxes without server setup.

Human-in-the-loop multi-agent orchestration from browser and mobile.

Confirmed Assumptions

• Agent Board (multi-agent orchestration) is the core value. [from prompt]
• Hybrid SaaS + optional managed IaaS is the best path. [from prompt]
• Zero-inbound security posture is a genuine differentiator. [from prompt]
• BYO-client SaaS should come before full managed IaaS. [from prompt]

Unknowns Requiring Validation

Ready for ICP once /pack install business-discovery is run.

ICP Inputs

• Priority segment: power users / solo founders (starting hypothesis, to be validated).
• Value wedge: always-on agent workstation, not browser IDE.
• Pricing: genuinely unknown — deferred to monetization research.
• Differentiators vs. T3 Code, browser IDEs, raw VPS+tmux.

Assumptions to Test First

1. Codex CLI programmatic control feasibility (technical spike).
2. Whether target users currently run multiple AI agents in parallel.
3. Whether zero-inbound security posture matters or is table stakes.

Resolved sequencing: Business discovery first.

1. /pack install business-discovery — install ICP, competitive analysis, value prop, positioning, lean canvas.
2. /icp — starting with power users / solo founders hypothesis.
3. Monetization research for pricing (no anchoring on preliminary ranges).
4. Technical spike on Codex CLI integration (after business discovery).