Yaowei Zheng
d4faee3a1e
Branch-length batch covering tooling, the model layer, the Web App and the public surfaces. Highlights: - Changelog: a per-release `changelog/<version>/` tree, grouped by the surface each change touches, with a root CHANGELOG.md holding one line per release. - Dev startup: `scripts/dev-prebuild.mjs` serializes the skills+core prebuild behind a lock and keeps `pnpm install` current; `pnpm dev` runs server+web together. - AgentHub 0.3.3 -> 0.4.0: OmniMessage complete payloads carry one opaque `fidelity` object in place of item-level `signature`/`phase`, threaded verbatim through Trace, replay and resume; malformed classification adapted to the new error types. - Model layer: a model is always referenced by an explicit `(provider, model_id)` pair. The provider is never inferred, guessed or defaulted -- both the catalog inference and the unique-match config resolution are gone, and CLI, SDK, server routes and run_subagent all require the complete pair. Catalog gains the Qwen Token Plan, Qwen Pay-As-You-Go and Fireworks AI gateways, plus an expanded OpenRouter group. - Web App: catalog preset sync and per-group speed test on the Models page, positional slash commands, a markdown renderer, skill-library update reminders, and a vertically centred draft page whose upward menus size themselves to the room available. - Public surfaces: restructured READMEs, the penguin.ooo landing site and blog, refreshed benchmark results for both suites, and the demo videos playing on the landing page. Includes the fixes from a full review of the branch: 23 confirmed findings, among them a provider-inference bug that could send one vendor's API key to another vendor's endpoint, and an Escape handler that destroyed the composer's contents unrecoverably. Verified on the branch head: pnpm test (1127 passing, 7 packages), pnpm typecheck and pnpm format:check clean, Playwright e2e 14/14.
@prismshadow/penguin-web
The PenguinHarness Web App — a React 19 + Vite + Tailwind CSS 4 SPA that renders the OmniMessage stream (same protocol and statistics as the CLI) and manages Agents, Skills, Models, usage and Traces. Feature tour: Web App Guide.
Layout
src/
├── main.tsx / app.tsx / router.tsx / styles.css
├── api/ # fetch wrapper, typed endpoint functions, EventSource (SSE) wrapper
├── state/ # auth / project / sessions / theme / locale contexts
├── lib/
│ ├── omni/ # OmniMessage stream → view-model reducer + connect-first/dedup controller
│ └── … # formatting, i18n dictionaries (zh/en), attachments, helpers
├── components/ # ui primitives (modal, drawer, select, …) + app layout
└── features/ # chat / agents / skills / models / usage / traces / benchmark / admin
DTO types are imported type-only from @prismshadow/penguin-server/api; no server code enters the bundle. Rendering rules for streaming partials (start/delta/stop aggregation, complete-message replacement, origin-chain nesting into subagent cards) live in lib/omni/stream-model.ts, which is fully unit-tested.
Development
Prereqs: Node >= 24, pnpm; run pnpm install at the repo root first (core must be built — the root dev:* scripts handle that).
pnpm dev:server # backend at 127.0.0.1:7364
pnpm dev:web # Vite dev server at 127.0.0.1:7365; /api proxied (SSE passes through)
The proxy target defaults to http://127.0.0.1:7364 (PENGUIN_API_PROXY overrides). Auth is a same-origin HttpOnly cookie, so the proxy keeps everything same-origin.
pnpm --filter @prismshadow/penguin-web typecheck
pnpm --filter @prismshadow/penguin-web test # vitest (pure logic)
pnpm --filter @prismshadow/penguin-web test:e2e # Playwright against a mock LLM
pnpm --filter @prismshadow/penguin-web build # vite build → dist/
Production
No separate static server needed: @prismshadow/penguin-server auto-hosts packages/web/dist (or PENGUIN_WEB_DIST) with an SPA fallback — build the web app, start the server, done. The published npm packages bundle the built front end.
Part of PenguinHarness · Apache-2.0