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-core
The PenguinHarness SDK and execution engine: the ReAct loop (context_engine), the OmniMessage protocol, the LLM / Environment interface contracts, Agent State and append-only Traces.
The engine speaks only OmniMessage and delegates everything else through two swappable interfaces — LLMInterface (models, via the @prismshadow/agenthub gateway) and EnvironmentInterface (tool execution). The SDK caller is the Human boundary: one entry point, session.run, streams the whole loop.
import { createAgent, isCompleteModelMessage, userText } from "@prismshadow/penguin-core";
const agent = await createAgent({ agentId: "default_agent" });
const session = await agent.createSession({ workspaceDir: process.cwd() });
for await (const output of session.run([userText("Create hello.txt containing hi")], {
approve: async () => "allow", // per-tool-call approval
})) {
if (isCompleteModelMessage(output) && output.payload.type === "text") {
console.log(output.payload.text);
}
}
A single run drives a complete Task: streaming output, per-call approvals, concurrent tool execution, interrupt carry-over, automatic reconnect and context compaction. State lives under ~/.penguin/data (PENGUIN_HOME); every Session restores fully from its Trace.
Documentation
Development
pnpm --filter @prismshadow/penguin-core build # tsup → dist/ (exports point at dist)
pnpm --filter @prismshadow/penguin-core typecheck
pnpm --filter @prismshadow/penguin-core test
pnpm test:e2e # live-model e2e (needs DEEPSEEK_API_KEY)
Part of PenguinHarness · Apache-2.0