Skip to content

Polish Frontier submission docs #172

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
nullxnothing wants to merge 3 commits into
base: main
Choose a base branch
from
Open

Polish Frontier submission docs #172

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
3b9292f
docs: sharpen Frontier judging narrative
nullxnothing May 20, 2026
1f3111a
docs: make Frontier submission self-contained
nullxnothing May 20, 2026
b6c9135
security: confirm registry state changes
nullxnothing May 20, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
80 changes: 58 additions & 22 deletions FRONTIER_SUBMISSION.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,25 +2,42 @@

DAEMON is a Solana-native agent workbench for verifiable AI development work.

The product turns agent coding sessions into auditable work: a developer opens a project, enables the Solana and Helius MCP stack, spawns an agent, funds a task, receives a work receipt, and approves settlement on devnet. The on-chain registry makes the relationship between prompt, task, agent session, receipt, approval, and settlement inspectable instead of relying on chat logs.
The product turns agent coding sessions into auditable work: a developer opens a project, enables the Solana and Helius MCP stack, spawns an agent, creates and funds a task, receives a work receipt, reviews it, and approves settlement on devnet. The on-chain registry makes the relationship between prompt, task, agent session, receipt, approval, and settlement inspectable instead of relying on chat logs.

## Positioning

**Short pitch:** Verifiable AI dev work for Solana teams.

**Category:** Agent workbench, not a generic IDE.

**Audience:** Solana founders, protocol teams, and independent builders who already use coding agents but need repeatable project context, MCP setup, wallet/RPC readiness, and proof that paid agent work was requested, delivered, reviewed, and settled.
**Audience:** Solana founders, protocol teams, agencies, and independent builders who already use coding agents but need repeatable project context, MCP setup, wallet/RPC readiness, and proof that paid agent work was requested, delivered, reviewed, and settled.

**Why Solana:** Solana teams already operate through wallets, RPCs, explorers, devnet/mainnet deploy flows, and on-chain settlement. DAEMON puts the agent workflow next to those primitives instead of leaving the proof trail scattered across terminal logs and chat transcripts.

## Canonical Demo

**One sentence:** DAEMON turns an AI coding task into a devnet-verifiable work receipt and settlement trail.

The demo should show one uninterrupted flow:

1. Open a Solana project in DAEMON.
2. Enable Solana/Helius MCP readiness.
3. Spawn Claude or Codex in a dedicated PTY tab.
4. Create a devnet agent work task with prompt, acceptance criteria, bounty, owner, verifier, and agent wallet.
5. Fund/start the task.
6. Submit receipt hashes for commit, diff, tests, and artifact URI.
7. Approve and settle the task.
8. Open the Solana Explorer devnet transaction proof.

## 2-Minute Demo Runbook

| Time | Scene | What to show |
| --- | --- | --- |
| 0:00-0:15 | Open project | Launch DAEMON, open `C:\Users\offic\Projects\DAEMON`, show project tabs, editor, and terminal. |
| 0:00-0:15 | Open project | Launch DAEMON, open a clean Solana project, show project tabs, editor, and terminal. |
| 0:15-0:30 | Enable Solana/Helius MCP | Open Project Readiness, toggle Solana/Helius MCP, confirm wallet, RPC, and MCP readiness checks. |
| 0:30-0:50 | Spawn agent | Open Agent Launcher, choose Claude or Codex, attach Solana/Helius MCP, spawn an agent in a dedicated PTY tab. |
| 0:50-1:10 | Create and fund task | In Agent Work or Registry controls, create a devnet task with bounty/escrow funding and show the task account. |
| 1:10-1:25 | Submit work receipt | Have the agent submit a receipt hash/summary for the completed work and show the receipt state. |
| 1:10-1:25 | Submit work receipt | Submit receipt hashes/summary for completed work and show the receipt state. |
| 1:25-1:45 | Approve and settle | Approve the receipt, settle the task on devnet, and show the final state transition. |
| 1:45-2:00 | Explorer proof | Open the Solana Explorer devnet transaction link and close on the value prop: verifiable AI dev work for Solana teams. |

Expand All @@ -29,39 +46,58 @@ Recording checklist:
- Keep the capture under 2 minutes.
- Use a clean project with funded devnet wallet and `RPC_URL` set.
- Copy the final explorer URL into the submission notes.
- Mention the test status: `pnpm run typecheck`, `pnpm test`, and `cargo test` pass locally.
- To capture the final take with `ffmpeg`, run:
- Mention the local validation status: `pnpm run typecheck`, `pnpm test`, and Anchor/Cargo registry tests.
- Do not lead with token/community materials; lead with the working devnet proof flow.
- To capture a 120 second take with `ffmpeg`, run:

```powershell
powershell -ExecutionPolicy Bypass -File scripts/frontier-record.ps1
```

The helper records the desktop for 120 seconds to `docs/frontier-demo.mp4`.
The helper records the desktop to `docs/frontier-demo.mp4`.

Devnet preflight:
## Devnet Preflight

- Confirm the registry program is current: `solana program show 3nu6sppjDtAKNoBbUAhvFJ35B2JsxpRY6G4Cg72MCJRc --url devnet`.
- If the deployed binary is stale, rebuild and upgrade the registry before recording: `cargo build-sbf` from `programs/daemon-registry`, then `solana program deploy --url devnet --program-id 3nu6sppjDtAKNoBbUAhvFJ35B2JsxpRY6G4Cg72MCJRc programs/daemon-registry/target/sbpf-solana-solana/release/daemon_registry.so`.
- Confirm the registry program is current:

## Business Model
```bash
solana program show 3nu6sppjDtAKNoBbUAhvFJ35B2JsxpRY6G4Cg72MCJRc --url devnet
```

DAEMON can monetize through four explicit channels:
- If the deployed binary is stale, rebuild and upgrade the registry before recording:

1. **Pro desktop:** Paid local app tier for advanced agent orchestration, session replay, telemetry, registry publishing, and premium Solana tooling.
2. **Team workspaces:** Shared project profiles, policy-controlled MCP configs, audit logs, agent session history, and role-based approval flows for protocol teams.
3. **Paid MCP/plugin marketplace:** Curated Solana protocol packs, MCP servers, agent skills, and workflow plugins with creator revenue share.
4. **Agent-work escrow fees:** A small protocol or platform fee on task funding and settlement when teams pay agents or contractors through the on-chain registry.
```bash
cd programs/daemon-registry
cargo build-sbf
solana program deploy --url devnet --program-id 3nu6sppjDtAKNoBbUAhvFJ35B2JsxpRY6G4Cg72MCJRc target/sbpf-solana-solana/release/daemon_registry.so
```

- Use a devnet wallet with enough SOL for task funding, rent, and fees.
- Save the final explorer link beside the demo video for judges.

## Frontier Criteria Fit

**Functionality and code quality:** Electron app, typed IPC bridge, PTY-backed agent sessions, project-level MCP management, wallet/RPC readiness, and Anchor registry tests.
| Criterion | DAEMON answer |
| --- | --- |
| Functionality and code quality | Electron app, typed IPC bridge, PTY-backed agent sessions, project-level MCP management, wallet/RPC readiness, SQLite migrations, smoke tests, CI, and Anchor registry code. |
| Impact | Solana builders are adopting coding agents, but teams still need a trustworthy way to request, review, verify, and settle agent work. |
| Novelty | The core wedge is not another editor. It is a workbench where AI development tasks can produce devnet-verifiable receipts and settlements. |
| Solana UX | Solana/Helius MCP setup, wallet readiness, RPC checks, devnet registry transactions, task funding, and explorer links are first-class flows. |
| Open source and composability | MIT-licensed repo with an extensible plugin/MCP architecture and an Anchor program that other tools can inspect or build against. |
| Business plan | Pro desktop, team workspaces, marketplace fees, and agent-work escrow fees create a path from individual builders to protocol/team usage. |

**Impact:** Solana builders are already adopting coding agents; DAEMON targets the missing trust layer between autonomous work, human approval, and payment.
## Business Model

**Novelty:** The core wedge is not another editor. It is a workbench where AI development tasks can produce devnet-verifiable receipts and settlements.
DAEMON can monetize through four explicit channels:

**Solana UX:** Solana/Helius MCP setup, wallet readiness, RPC checks, devnet registry transactions, and explorer links are first-class user flows.
1. **Pro desktop:** Paid local app tier for advanced agent orchestration, session replay, telemetry, registry publishing, and premium Solana tooling.
2. **Team workspaces:** Shared project profiles, policy-controlled MCP configs, audit logs, agent session history, and role-based approval flows for protocol teams.
3. **Paid MCP/plugin marketplace:** Curated Solana protocol packs, MCP servers, agent skills, and workflow plugins with creator revenue share.
4. **Agent-work escrow fees:** A small protocol or platform fee on task funding and settlement when teams pay agents or contractors through the on-chain registry.

**Open source and composability:** MIT-licensed repo with an extensible plugin/MCP architecture and an Anchor program that other tools can inspect or build against.
## Judge Notes

**Business plan:** Pro desktop, team workspaces, marketplace fees, and agent-work escrow fees create a path from individual builders to paid protocol/team usage.
- DAEMON is not a VS Code fork.
- The demo should be evaluated as a Solana-native agent workbench and receipt layer.
- The important proof is the task lifecycle: prompt -> task -> agent session -> receipt -> approval -> settlement -> explorer link.
- Community/token information is intentionally separated from the core product narrative so the submission focuses on functionality, Solana UX, and composability.
50 changes: 35 additions & 15 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,9 +13,9 @@

<p align="center">
<a href="https://daemon-landing.vercel.app">Website</a> &middot;
<a href="#frontier-hackathon">Frontier</a> &middot;
<a href="#install">Install</a> &middot;
<a href="#features">Features</a> &middot;
<a href="FRONTIER_SUBMISSION.md">Frontier</a> &middot;
<a href="#architecture">Architecture</a> &middot;
<a href="#development">Development</a> &middot;
<a href="CONTRIBUTING.md">Contributing</a>
Expand All @@ -24,26 +24,39 @@
---

<p align="center">
<strong>$DAEMON CA:</strong> <code>4vpf4qNtNVkvz2dm5qL2mT6jBXH9gDY8qH2QsHN5pump</code>
<img src="docs/screenshots/ui-overview.webp" alt="DAEMON agent workbench with editor, terminal, and sidebar" width="800">
</p>

---
DAEMON is a standalone Electron workbench for Solana developers who use autonomous coding agents. It combines an offline editor, real PTY terminals, Claude/Codex agent spawning, MCP server management, wallet/RPC readiness, and an Anchor-backed registry for publishing verifiable agent work receipts on devnet.

<p align="center">
<img src="docs/screenshots/ui-overview.webp" alt="DAEMON agent workbench with editor, terminal, and sidebar" width="800">
</p>
The wedge is simple: AI agents can write code, but Solana teams still need proof of what was requested, what was delivered, who reviewed it, and how settlement happened.

## Frontier Hackathon

**[Frontier demo runbook](FRONTIER_SUBMISSION.md#2-minute-demo-runbook)** — 2-minute submission flow from project open to devnet settlement.
**Short pitch:** verifiable AI dev work for Solana teams.

DAEMON is a standalone Electron agent workbench for Solana developers who use autonomous coding agents. It combines an offline editor, real PTY terminals, Claude/Codex agent spawning, MCP server management, wallet/RPC readiness, and an Anchor-backed registry for publishing verifiable agent work receipts on devnet. Not a VS Code fork.
**Demo path:** open project -> enable Solana/Helius MCP -> spawn Claude or Codex -> create and fund a devnet task -> submit receipt hashes -> approve and settle -> open Solana Explorer proof.

See the full [Frontier submission runbook](FRONTIER_SUBMISSION.md#2-minute-demo-runbook) for the 2-minute recording flow and devnet preflight.

### Judging checklist

| Criterion | DAEMON fit |
| --- | --- |
| Functionality and code quality | Electron app with typed IPC, PTY-backed agent sessions, project MCP management, wallet/RPC readiness, CI, smoke tests, and an Anchor registry. |
| Impact | Solana builders already use AI coding agents; DAEMON adds the missing trust layer for requested, delivered, reviewed, and settled work. |
| Novelty | Agent sessions produce task, receipt, review, and settlement evidence instead of relying on chat logs. |
| Solana UX | Wallet readiness, Solana/Helius MCP setup, devnet task funding, registry transactions, and explorer links are first-class flows. |
| Open source and composability | MIT-licensed app plus an Anchor registry program that other tools can inspect or build against. |
| Business plan | Pro desktop, team workspaces, paid MCP/plugin marketplace, and agent-work escrow fees. |

## Install

**Windows:** Download the [latest .exe](https://github.com/nullxnothing/daemon/releases/latest/download/DAEMON-setup.exe)

<a name="mac-install"></a>

**Mac:** Build from source (signed builds configurable via Apple credentials):
**Mac:** Build from source. Signed builds are configurable with Apple Developer credentials:

```bash
git clone https://github.com/nullxnothing/daemon.git
Expand All @@ -53,11 +66,11 @@ pnpm run build
pnpm run package
```

The `.dmg` will be in `release/2.0.0/`. Signed/notarized builds require Apple Developer credentials in the packaging environment. Without them, the app will still package, but Gatekeeper may require right-click > Open on first launch.
The `.dmg` is emitted under the current release folder, for example `release/3.0.13/`. Without signing credentials, macOS may require right-click > Open on first launch.

<a name="linux-install"></a>

**Linux:** Build from source (AppImage builds coming soon):
**Linux:** Build from source. AppImage builds are supported through the package script:

```bash
git clone https://github.com/nullxnothing/daemon.git
Expand All @@ -67,7 +80,7 @@ pnpm run build
pnpm run package
```

The AppImage will be in `release/2.0.0/`. Make it executable with `chmod +x` and run directly.
The AppImage is emitted under the current release folder, for example `release/3.0.13/`. Make it executable with `chmod +x` and run directly.

**Build from source (any platform):**

Expand All @@ -94,7 +107,9 @@ Requires **Node.js 22+** and **pnpm 9+**.
<img src="docs/screenshots/agent-launcher.webp" alt="Agent launcher with model selection and MCP config" width="800">
</p>

**Agent Launcher** — Spawn Claude Code agents with custom system prompts, model selection, and per-project MCP configurations. Agents run as real CLI sessions in dedicated terminal tabs.
**Agent Launcher** — Spawn Claude Code or Codex agents with custom system prompts, model selection, and per-project MCP configurations. Agents run as real CLI sessions in dedicated terminal tabs.

**Agent Work Registry** — Create, fund, start, submit, approve, reject, settle, or expire agent work tasks with hashes for the repo, prompt, acceptance criteria, diffs, tests, and artifacts.

**MCP Management** — Toggle project-level and global MCP servers from the sidebar. Changes write directly to `.claude/settings.json` and `.mcp.json` with a restart indicator when configs change.

Expand All @@ -112,7 +127,7 @@ Requires **Node.js 22+** and **pnpm 9+**.

**Embedded Browser** — Built-in browser with a security sandbox for previewing and testing.

**PumpFun Integration** — Token launches and bonding curve interactions directly from the IDE.
**Solana Integrations** — Wallet, launch, SpawnAgents, PumpFun, x402/MPP, and protocol workflow surfaces inside one local workbench.

**Multi-Project Tabs** — Tabbed project switching with per-project terminal sessions, MCP configs, and file trees. Context switching without losing state.

Expand Down Expand Up @@ -156,6 +171,7 @@ Key decisions:
| Database | better-sqlite3 (WAL) |
| Git | simple-git |
| Packaging | electron-builder |
| Registry | Anchor / Solana devnet |

## Development

Expand All @@ -165,11 +181,15 @@ pnpm run dev # Dev server with hot reload
pnpm run typecheck # TypeScript validation
pnpm run test # Run tests (Vitest)
pnpm run build # Production build
pnpm run package # Create distributable (.exe / .dmg)
pnpm run package # Create distributable (.exe / .dmg / AppImage)
```

See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines on pull requests and code style.

## Community

$DAEMON community token CA: `4vpf4qNtNVkvz2dm5qL2mT6jBXH9gDY8qH2QsHN5pump`

## License

[MIT](LICENSE)
Loading
Loading