Token导航 LogoToken导航TokenDH.com
开发需要联网github未标认证来源可访问许可证需确认审计提醒

codex-upcoming-featuresCodex upcoming features 命令行

Agent Skill

codex-upcoming-features 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

524

周安装

21

GitHub Stars

53

下载量

170
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:codex-upcoming-features(Codex upcoming features 命令行)
来源仓库:https://github.com/tkersey/dotfiles
仓库路径:skills/codex-upcoming-features
安装命令:
npx skills add https://github.com/tkersey/dotfiles --skill codex-upcoming-features
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/tkersey/dotfiles --skill codex-upcoming-features

简介

codex-upcoming-features 同步 openai/codex 仓库并提取未发布功能摘要。

  • 主要依据源码中的 feature stages 与元数据生成报告。
  • 需先验证 gh auth status 并通过本地克隆路径解析。
  • 输出包含 commit/PR 分析的补充证据链说明。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Codex Upcoming Features

Overview

Sync a durable local clone of openai/codex and produce an exhaustive summary of unreleased features and notable improvements since the latest stable (non-prerelease) release tag.

Primary evidence comes from source files that define feature stages and user-facing metadata. Commit/PR analysis is secondary supporting context.

Workflow

  1. Run gh auth status and fail closed if GitHub CLI is not authenticated.
  2. Resolve durable local clone path:
  • default: $CODEX_HOME/repos/<owner>-<repo> (fallback: ~/.codex/repos/<owner>-<repo>)
  • override: --local-repo <path>
  1. If clone is missing, clone from https://github.com/<owner>/<repo>.git.
  2. Sync local clone on every run (git fetch --tags --prune origin, checkout default branch, git pull --ff-only).
  3. Mine these source files on every run (fail closed if missing):
  • codex-rs/core/src/features.rs
  • codex-rs/tui/src/chatwidget.rs
  • codex-rs/core/src/codex.rs
  • codex-rs/app-server/src/codex_message_processor.rs
  • codex-rs/app-server-protocol/src/protocol/v2.rs
  • codex-rs/app-server-protocol/schema/json/v2/ExperimentalFeatureListResponse.json
  • codex-rs/app-server/README.md
  • codex-rs/tui/src/tooltips.rs
  • announcement_tip.toml
  1. Extract source-derived upcoming feature data from features.rs:
  • Stage::Experimental (beta/user-facing now)
  • Stage::UnderDevelopment (upcoming but not broadly ready)
  • include name, menu_description, and announcement where available
  1. Extract structured supporting evidence from each other mined source file (UI wiring, app-server mapping, protocol enum/schema, docs contract, tooltip/announcement pipeline).
  2. Record analyzed commit SHA from synced local clone (git rev-parse HEAD) and include it in output.
  3. Resolve repo default branch and latest stable release tag (exclude draft/prerelease).
  4. Compute unreleased range from local git history: stable_tag..origin/default_branch.
  5. Classify each unreleased commit into one of:
  • feature
  • notable_improvement
  1. Exclude entries that are clearly non-user-impacting (for example docs/chore/test-only).
  2. Emit an exhaustive report with:
  • UTC as-of timestamp
  • local repo path used
  • analyzed commit SHA
  • stable baseline tag and compare link
  • source files mined
  • source-derived feature list (primary)
  • cross-file source evidence (primary)
  • per-item PR/commit evidence links

Commands

Run markdown summary (default):

uv run python scripts/summarize_upcoming.py --repo openai/codex

Run JSON output:

uv run python scripts/summarize_upcoming.py --repo openai/codex --output json

Override baseline tag if needed:

uv run python scripts/summarize_upcoming.py --repo openai/codex --base-tag rust-v0.106.0

Override durable local clone path:

uv run python scripts/summarize_upcoming.py --repo openai/codex --local-repo ~/workspace/cache/openai-codex

Output Contract

Always include:

  • As of (UTC) timestamp
  • local repo path used for summarization
  • analyzed commit SHA
  • Stable baseline tag
  • compare URL
  • source files mined
  • source-derived upcoming feature sections (beta + underDevelopment)
  • source-derived cross-file evidence section covering all required files
  • exhaustive list of included items
  • source links for each item

When there is a conflict between source-derived feature data and commit text, source-derived data wins.

JSON mode returns:

  • as_of_utc
  • repo
  • local_repo
  • analyzed_commit_sha
  • stable_tag
  • compare_url
  • source_files_mined[]
  • source_primary.beta_features[] and source_primary.under_development_features[]
  • source_primary.supporting_evidence (structured extraction for every required source file)
  • items[] with kind, title, commit_sha, commit_url, pr_number, pr_url (secondary)

Classification Rules

Read references/classification-rules.md before changing filtering behavior. Keep inclusion deterministic and biased toward user-visible changes.

Guardrails

  • Do not use stale local commit snapshots; clone if missing and sync (fetch/pull) on every run.
  • Fail closed if local repo origin does not match the requested owner/repo.
  • Keep release resolution live (GitHub API), source feature extraction local (from mined files), and commit extraction local (synced git history).
  • Treat commit messages as supporting evidence only; never treat commits as primary feature truth when source files disagree.
  • Do not infer future roadmap from open issues or open PRs in this skill version.
  • Fail closed with actionable errors when API/auth/rate-limit calls fail.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

34.64%
按下载量换算59

Claude

29.73%
按下载量换算51

Cursor

17.15%
按下载量换算29

Gemini CLI

9.47%
按下载量换算16

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills