本文へ移動
cccskills
無料GitHub で公開

guardrails

How AgentGuards enforces its guardrails in Codex — fully automatic, via hooks, with nothing for you to call. Load this if asked how AgentGuards works, whether you need to screen something yourself, or why no agentguards check_input / authorize_action tools are available.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md1.8 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。

AgentGuards — enforced by hooks

This plugin has no MCP server, so there is no check_input, authorize_action or health_check tool — do not look for them, they are not installed.

That is not reduced protection. The hooks in hooks/hooks.json enforce deterministically, on every request, whatever the model does:

  • UserPromptSubmit screens every prompt before you see it.
  • PreToolUse authorizes every shell command before it runs: denied, or allowed, or left to Codex's own approval prompt.
  • PermissionRequest rides that approval prompt: it denies what AgentGuards rejects and remembers what the user approved for the rest of the session.
  • PostToolUse scans the output of shell commands that fetch (curl, wget, http, fetch, aria2c — also behind sudo, timeout, bash -c, $(...)), withholding or redacting it before you read it, and scans apply_patch edits for vulnerabilities and secrets.

Nothing here requires you to call a tool, check a decision, or format a block message — the hook has already acted by the time you see (or don't see) the result. If a hook blocks something, reply with its message rather than composing your own.

One limit, stated plainly: Codex's built-in web search runs on OpenAI's servers, not on this machine, so no hook ever sees its results. Content fetched with a shell command is screened; built-in web search results are not.

To check that the guardrails are on, use the status skill. To set the API key, use the setup skill.

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-plugins1,2752026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-plugins1,2752026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-plugins1,2752026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-plugins1,2752026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-plugins1,2752026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-plugins1,2752026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