本文へ移動
cccskills
無料GitHub で公開

agy-native

Run a supplied task in headless AGY (Antigravity, Gemini) and collect its result. Use when: AGY, Antigravity or Gemini is requested by name; never a fallback.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md3.6 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

AGY Native

Run a supplied packet in AGY (Antigravity) only when the caller explicitly selects that runtime. Scope the run to the supplied workspace and packet, return runtime evidence, and stop.

Before launch

  1. Read the live surface. Run agy --help for flags and defaults and agy models for the current model set. The CLI changes faster than skill text. Wrapper drift is a run built from remembered syntax that silently changed: it looks scoped and is not.
  2. Absent means stop. If agy is missing or its help cannot be read, report the absence and stop. Never fall back silently to Codex, Claude or any other runtime; a substitute needs a new caller selection. Nothing routes work to AGY on its own either. The repository reviewer library rates AGY a routine-tier, degraded-fallback reviewer; that rating applies only when a caller opts in, and it is not an automatic route.
  3. Set an explicit bound. Print mode (agy -p / --print) is the headless path, and it has no built-in limit: --print-timeout defaults to 0, which waits until the turn completes (AGY 1.2.14). Pass --print-timeout from the caller's remaining deadline and hold the same bound outside the process. A run stopped at the bound is a timeout, not a result.
  4. Name the posture (below) and confirm the packet's declared effects and the caller's authorization cover it.
  5. Keep author and validator apart. A validator is a new run in a new conversation, never -c/--continue or --conversation <author-id>. A session that reviews its own work forfeits the fresh judgment that makes a validator's evidence usable. Validators stay read-only and hand judgment to Validate; they never write the core verdict.

Permission posture (disclose it; never assume it)

Flags choose the posture. Name the one used:

  • No posture flag: AGY prompts for each tool permission. A headless run has no one to answer, so a prompt can stall it until the bound.
  • --dangerously-skip-permissions: auto-approves every tool call. Use it only when the declared effects and the caller's authorization cover that blast radius.
  • --sandbox: restricts the session's terminal access.
  • --mode (plan, accept-edits): changes the execution mode; confirm its current meaning in agy --help.

Return

command:      exact argv (prompt by reference)
posture:      permission, sandbox and mode flags used
model:        requested model; observed model if AGY reported one, else unknown
conversation: id AGY reported, else unknown
bound:        --print-timeout value and the outer deadline
outcome:      exit status | timed out | not run (reason)
artifacts:    captured output and changed-file paths

AGY plugin, memory, permission, retry and session state are runtime facts; they never become AgentOps phase, queue or completion state. Installation, plugin mutation and recurring scheduling need separate explicit authorization.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Use AtomLane to compile and execute safe atomic parallel plans on macOS and native Windows Preview for worthwhile independent argv tasks, dependency DAGs, supported platform entrypoints, or Apple-silicon operators. Use at task start or an execution boundary when structured local work may contain two or more worthwhile units; skip plain answers, one quick command, and work whose effects cannot be safely bounded.

日本語の概要は準備中です。原文の説明を表示しています。

hashgraph-online/awesome-codex-plugins1,2832026年10月11日 更新

add

無料

Register a deferred decision in the debt registry. Trigger by judgment, not a marker scan, whenever a future reader would ask "why this way?": an unmade decision, stub, loosened type, bypassed check, swallowed error, a default picked "for now", or a TODO/FIXME/HACK/XXX marker. Trigger immediately whenever you defer work, or when the user invokes $add. Over-register freely; the developer drops with "drop A", "drop A,C", or "drop all".

日本語の概要は準備中です。原文の説明を表示しています。

hashgraph-online/awesome-codex-plugins1,2832026年10月11日 更新

ADK 框架适配层。为 LangChain / EINO / AutoGen / AgentScope / CrewAI 提供框架特定的 代码模板、惯用模式、API 映射和项目结构,供 agent-dev-workshop Phase 5 代码生成使用。 每个框架 reference 文件标注 verified_date 用于版本锁定。

日本語の概要は準備中です。原文の説明を表示しています。

hashgraph-online/awesome-codex-plugins1,2832026年10月11日 更新

中文调试修复技能。用于报错、测试失败、页面异常、功能不符合预期、需要定位根因并做最小修复时。触发语包括"进入调试模式""帮我修问题""报错了""测试失败""页面坏了""找根因"。

日本語の概要は準備中です。原文の説明を表示しています。

hashgraph-online/awesome-codex-plugins1,2832026年10月11日 更新

交互式 AI Agent 开发工作坊:通过 6 阶段深度协作对话,引导用户完成 Agent 需求分析、架构设计、 工具定义、Prompt 与编排设计、代码生成、验证迭代,产出可直接运行的 Agent 项目。 框架无关设计优先,支持 LangChain / EINO / AutoGen / AgentScope / CrewAI 等 ADK 框架。

日本語の概要は準備中です。原文の説明を表示しています。

hashgraph-online/awesome-codex-plugins1,2832026年10月11日 更新

中文漂移审计技能。用于项目或学习过程变乱、上下文漂移、任务分叉、多个方案冲突、命名不一致、Codex 可能顺手改多了时。触发语包括"漂移检查""感觉跑偏了""项目变乱了""检查是否失控""分叉太多""上下文漂移"。

日本語の概要は準備中です。原文の説明を表示しています。

hashgraph-online/awesome-codex-plugins1,2832026年10月11日 更新

hashgraph-online のスキルをすべて見る

このスキルの問題を報告する