External/buzz

Fork 0

mirror of https://github.com/block/buzz.git synced 2026-08-18 06:50:31 +02:00

Files

T

Will PflegerandGitHub 6dcb04bcd9 Fix CLI output normalization and wrong Nostr kinds (MCP parity) (#612 )

2026-05-22 10:32:20 -04:00

10 KiB

Raw Blame History

AGENTS.md — AI Agent Contributor Guide

This guide is for AI agents contributing to the Sprout codebase. It covers agent-specific context and conventions. For general contributor info (setup, code style, PR process, architecture), see CONTRIBUTING.md.

Repo Structure

crates/
  sprout-relay        # WebSocket relay server — main entry point
  sprout-core         # Core types, event verification, filter matching
  sprout-db           # Postgres event store and data access layer
  sprout-auth         # Authentication and authorization
  sprout-pubsub       # Redis pub/sub fan-out, presence, typing indicators
  sprout-mcp          # MCP server providing AI agent tools
  sprout-acp          # ACP harness bridging Sprout events to AI agents
  sprout-workflow     # YAML-as-code workflow engine (evalexpr conditions)
  sprout-search       # Typesense-backed full-text search
  sprout-audit        # Hash-chain audit log
  sprout-huddle       # LiveKit audio/video integration
  sprout-proxy        # Nostr client compatibility proxy
  sprout-admin        # Operator CLI for relay administration
  sprout-test-client  # Integration test client and E2E test suite
  sprout-sdk          # Typed Nostr event builders (used by sprout-mcp and sprout-cli)
  sprout-media        # Blossom/S3 media storage
  sprout-cli          # Agent-first CLI

desktop/              # Tauri 2 + React 19 desktop app
migrations/           # SQL migrations (auto-applied on relay startup)
scripts/              # Dev tooling
.env.example          # Config template — copy to .env before running

Getting Started

. ./bin/activate-hermit   # activate hermit toolchain (Rust, Node, etc.)
cp .env.example .env      # configure local environment
just setup                # install deps, run migrations
just relay                # start relay at ws://localhost:3000
just ci                   # run before any PR

See CONTRIBUTING.md for full setup details and dependency requirements.

Quality Gates

Run just ci before every PR — it runs fmt + clippy + desktop lint + unit tests + builds. Clippy passing does not mean fmt passes; run both.

Run just test for integration tests if you touched sprout-relay, sprout-db, or sprout-auth — these require a running Postgres and Redis.

Additional rules:

No unsafe code
Do not introduce new unwrap() or expect() in production paths — use ? and proper error types
New public API must have doc comments

Key Patterns

Dual API surface: Sprout exposes both a REST API and a NIP-29 WebSocket relay. Both paths converge on shared DB functions in sprout-db. When adding a feature, implement the shared DB logic first, then wire up both surfaces.

Prefer Nostr events over new REST endpoints: For new feature work, model the operation as a Nostr event (new kind in sprout-core/src/kind.rs, handler in sprout-relay) rather than adding a new REST endpoint. REST is reserved for things that genuinely need an HTTP-only surface: media upload/download (Blossom), OAuth callbacks, health checks, and the existing read endpoints that proxy DB queries. Two helpful endpoints already exist and rarely need to be duplicated:

POST /events — submit any signed event (same path the WebSocket uses).
POST /query — Nostr REQ filters over HTTP. NIP-50 search filters are routed to sprout-search (Typesense-backed) automatically.
POST /count — Nostr COUNT filters over HTTP.

If you find yourself reaching for a new REST endpoint, first check whether an event kind would do the job — it usually will, and you get realtime fan-out, NIP-29 scoping, and the existing auth pipeline for free.

Reference https://github.com/nostr-protocol/nips

Event kinds: All event kind integers are defined in sprout-core/src/kind.rs. New features get new kind integers — add them here first, then implement handling in the relay.

Channel scoping: Channels use h tags (NIP-29 group tag), not e tags. Filters and queries must scope to h tags when operating within a channel.

MCP tools — dual transport: The MCP server in sprout-mcp uses two patterns: write operations send signed Nostr events over WebSocket; read operations call REST endpoints (see relay_client.rs for the HTTP helpers). Add the REST endpoint or event handler first, then add the MCP tool that calls it. Do not implement logic directly in MCP handlers.

Workflow conditions: sprout-workflow uses evalexpr for condition evaluation. Keep expressions simple and testable.

Thread counters: reply_count and descendant_count are materialized on thread root events. Any code that inserts replies must update these counters — check existing reply handlers for the pattern.

Agent CLI (`sprout-cli`)

sprout is the agent-first CLI replacing sprout-mcp. Auth env vars (SPROUT_RELAY_URL, SPROUT_PRIVATE_KEY) are auto-injected by the ACP harness into managed agent subprocesses.

All reads return sig-stripped JSON arrays; all writes return {event_id, accepted, message}; creates add the entity ID. Exit codes: 0=ok, 1=input error, 3=auth missing. See crates/sprout-cli/TESTING.md for the full live-testing runbook.

Testing

just test-unit    # unit tests, no infrastructure needed
just test         # full integration suite (requires Postgres + Redis)

E2E tests live in crates/sprout-test-client/tests/:

e2e_relay.rs — WebSocket relay protocol
e2e_rest_api.rs — REST endpoint coverage
e2e_mcp.rs — MCP tool surface
e2e_tokens.rs — auth token flows
e2e_workflows.rs — workflow engine
e2e_media.rs — media upload/download (Blossom)
e2e_media_extended.rs — extended media scenarios
e2e_nostr_interop.rs — Nostr interop (NIP-50 search, NIP-10 threads, NIP-17 gift wraps)

Desktop E2E: cd desktop && pnpm exec playwright test

See TESTING.md for the full multi-agent E2E guide.

Common Gotchas

Kind 39000 for channel metadata, not 41 — kind 41 is NIP-01 (unused). All kinds defined in sprout-core/src/kind.rs.
Relay queries must specify kinds — omitting kinds triggers the p-gate (403). Always include explicit kind filters.
Worktrees: cd in the same command — shell CWD doesn't persist between tool calls. Use cd /path && cargo build as one command.
Desktop fmt check fails in worktrees — run just desktop-tauri-fmt-check from the main checkout. CI is unaffected.

Desktop App

The desktop app is Tauri 2 + React 19 + Vite + Tailwind CSS. Features are organized under desktop/src/features/. Biome handles linting and formatting.

just desktop-dev   # web-only dev server (faster iteration)
just desktop-app   # full Tauri app with native shell

Workspace Switching

The desktop app supports multiple workspaces (each backed by a different relay). Switching workspaces does not reload the page — it uses React key-based remounting. <AppReady key={workspaceKey} /> in App.tsx forces the entire workspace-scoped subtree to unmount and remount with fresh state.

Module-level singletons must be explicitly reset. React remounting only clears React state (useState, useRef, context). Module-level variables (Maps, class instances, cached promises) survive across remounts. Every workspace-scoped singleton needs a reset function wired into resetWorkspaceState() in desktop/src/features/workspaces/useWorkspaceInit.ts.

Current singletons that are reset on workspace switch:

relayClient.disconnect() — WebSocket teardown + promise rejection
resetMediaCaches() — proxy port and relay origin caches
clearSearchHitEventCache() — search result event cache
clearAllDrafts() — message draft cache

If you add a new module-level cache, Map, or class instance that holds workspace-scoped data, you must add its reset to resetWorkspaceState(). Failure to do so causes data from the old workspace to leak into the new one.

Key files:

desktop/src/app/App.tsx — workspace key, init gate, remount boundary
desktop/src/features/workspaces/useWorkspaceInit.ts — resetWorkspaceState(), applies config to Tauri backend
desktop/src/features/workspaces/useWorkspaces.tsx — WorkspacesProvider context (shared state for App + AppShell)
desktop/src/main.tsx — provider hierarchy (QueryClientProvider > WorkspacesProvider > App)

Mobile App (Flutter)

The mobile app lives in mobile/ — a Flutter app using Riverpod + Hooks.

Architecture

State management: Riverpod + flutter_hooks (HookConsumerWidget)
Theme: Catppuccin Latte (light) / Macchiato (dark) — matches desktop
Features: Isolated under lib/features/, shared code in lib/shared/
Nostr models: lib/shared/relay/nostr_models.dart — event kinds must stay in sync with desktop/src/shared/constants/kinds.ts

Rules

NEVER use StatefulWidget — always use HookConsumerWidget or ConsumerWidget with flutter_hooks for local state.
NEVER run flutter run, flutter build, flutter clean, or flutter upgrade — only flutter test, flutter analyze, and dart format are safe for agents to run.
Do NOT use print() — use debugPrint() or structured logging.
Prefer context.colors and context.textTheme (via theme extensions) over raw Theme.of(context) calls.
Keep widgets small and composable.
Feature modules must not import from other feature modules — only from shared/.
Use Grid tokens for spacing, Radii for border radius.

Quality Checks

cd mobile
dart format --output=none --set-exit-if-changed .
flutter analyze
flutter test

Or from repo root: just mobile-check and just mobile-test.

Testing Conventions

Prefer widget tests over unit tests for UI components — test the whole widget tree, not individual methods.
Use ProviderScope(overrides: [...]) to inject fake notifiers.
Fake notifiers should extend the real notifier class and override build().
Use the WidgetHelpers.testable() wrapper for simple widget tests or build a custom ProviderScope + MaterialApp when you need specific overrides.

10 KiB Raw Blame History