Hacktoberfest 2026:維護者為十月標記出來的 issue,仍然開放、適合新手。 瀏覽 Hacktoberfest issue

[rig-summarizer] Daily file summary — 2026-10-19

未關閉
#681 0 則留言 0 個 reaction 已指派 0 人 在 GitHub 檢視

維護者通常 1 天內回覆

還沒有人認領這個 Issue。

評估

難度
1/5
預估耗時
1 小時以內
新手友好度
1/100
Issue 類型
文件
描述清晰度
需要釐清
活躍度
活躍
技術堆疊
typescript
領域
documentation

研究方向

This is an automatically generated daily file summary, not a request for a change. No target files, tests, or completion criteria are identified; first determine whether there is an actionable issue to address. The payload provides no specific work or definition of done.

由索引模型根據 Issue 內容生成。

描述

ai-agent automation
Global Summary

Rig is a minimal TypeScript agent harness. The core runtime (rig.ts) provides typed agents, prompt intents and a Copilot SDK runtime. Pluggable engines (Anthropic, Codex, DeepSeek, Gemini, Pi) share lifecycle helpers. A launcher and a launch tool run rig programs. The .github drivers and fixtures support a three-judge provider integration test, and the scripts directory holds live integration tests.

Note: this summary was produced directly from the first 20 .ts files, sorted alphabetically (the sub-agent runner was not used), so it is based on file headers and imports only.

Individual File Summaries
File Summary
./.agents/skills/rig/engines/anthropic.ts Anthropic engine for rig. It uses the Anthropic SDK with JSON-schema output and tool helpers.
./.agents/skills/rig/engines/codex.ts Codex engine factory built on @openai/codex-sdk, with optional thread options.
./.agents/skills/rig/engines/deepseek.ts DeepSeek engine factory wrapping DeepSeekHarness.
./.agents/skills/rig/engines/gemini.ts Gemini engine that spawns the Gemini CLI process, with configurable command, args, env and approval mode.
./.agents/skills/rig/engines/pi.ts Pi engine wrapping the Pi agent core and AI models, with tool support.
./.agents/skills/rig/engines/utils.ts Shared engine helpers: agentLifecycle (abort handling, rejects concurrent turns), objectToolSchema and toolResultText.
./.agents/skills/rig/globals.ts Ambient call, pipeline and parallel helpers that delegate to the active workflow run.
./.agents/skills/rig/launch-tool.ts Creates a Copilot SDK tool that launches rig programs through the launcher, with a timeout and URI/token validation.
./.agents/skills/rig/rig.ts Core runtime (~3.4k lines) with schemas (s.*), prompt intents (p.*), agents, addons, workflows, the Copilot SDK engine and the launcher CLI.
./.agents/skills/rig/run.ts Entry point for the installed skill; it dynamically imports rig.ts and runs the launcher CLI.
./.github/drivers/codex-fixture-mcp.ts Codex driver that serves the three-judge fixture launcher over a local HTTP MCP server.
./.github/drivers/copilot-sdk-driver.ts Copilot SDK driver that reads configuration from the environment and exposes the rig launch tool.
./.github/drivers/pi-rig-extension.ts Pi extension that provides a rig tool running the three-judge fixture and asserting its result.
./.github/fixtures/assert-three-judges.ts Asserts that the three-judge result has the right engine, model and judgments.
./.github/fixtures/pi-gateway.ts Loads and validates the Pi aw-gateway provider configuration with OpenAI-compatible APIs.
./.github/fixtures/provider-fixture.ts Selects the Codex, DeepSeek, Gemini or Pi engine from RIG_JUDGE_ENGINE and builds the three-judge fixture.
./.github/fixtures/provider-three-judges.ts Default export of the provider fixture.
./.github/fixtures/three-judges.ts Defines the three-judge workflow (clarity, safety, feasibility) over a dummy request.
./scripts/haiku.integration.test.ts Live integration tests running the haiku, sonnet and complex samples through the launcher when a token or SDK URI is set.
./scripts/rig-launch-tool.integration.test.ts Opt-in live test (RIG_TOOL_INTEGRATION=1) running the three-judge fixture through the harness-owned launch tool.

Generated by Daily Rig File Summarizer · copilot · auto · 14.9 AIC · ⌖ 0.52 AIC · ⊞ 7.6K · ◷

  • expires on Oct 26, 2026, 2:36 PM UTC
主要語言
TypeScript
星號
14
分支
0
平均合併
9 小時 49 分鐘
30 天內合併 PR
32

環境準備

這個專案沒有提供開發容器、Dockerfile 或貢獻指南,環境需要你自己搭建:先看它的 README,通用步驟見我們的新手貢獻指南。

從這裡開始

  1. 先讀完整個 Issue,再讀專案的貢獻指南。
  2. 在 Issue 下留言說明你要接手 —— 這能避免兩個人做同樣的事。
  3. Fork 儲存庫,在一個分支上完成修改。
  4. 送出 Pull Request,並在描述裡引用這個 Issue 編號。

githubnext/rig 的其他 Issue

查看 githubnext/rig 的全部 Issue

相似的 Issue

更多 TypeScript Issue

把新 issue 寄到你的電子郵件信箱

精選適合新手參與的 GitHub issue 摘要。