惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

F
Fortinet All Blogs
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
有赞技术团队
有赞技术团队
www.infosecurity-magazine.com
www.infosecurity-magazine.com
大猫的无限游戏
大猫的无限游戏
爱范儿
爱范儿
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
T
Threatpost
V
Visual Studio Blog
Apple Machine Learning Research
Apple Machine Learning Research
博客园 - Franky
人人都是产品经理
人人都是产品经理
cs.AI updates on arXiv.org
cs.AI updates on arXiv.org
The Cloudflare Blog
N
News and Events Feed by Topic
L
Lohrmann on Cybersecurity
cs.CL updates on arXiv.org
cs.CL updates on arXiv.org
酷 壳 – CoolShell
酷 壳 – CoolShell
V
V2EX
AWS News Blog
AWS News Blog
S
SegmentFault 最新的问题
T
Tailwind CSS Blog
Hugging Face - Blog
Hugging Face - Blog
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
Spread Privacy
Spread Privacy
J
Java Code Geeks
博客园 - 聂微东
T
Tor Project blog
宝玉的分享
宝玉的分享
博客园 - 叶小钗
Webroot Blog
Webroot Blog
博客园 - 【当耐特】
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
H
Heimdal Security Blog
Y
Y Combinator Blog
T
The Blog of Author Tim Ferriss
MongoDB | Blog
MongoDB | Blog
I
InfoQ
Security Latest
Security Latest
Martin Fowler
Martin Fowler
Hacker News: Ask HN
Hacker News: Ask HN
P
Privacy International News Feed
C
CERT Recently Published Vulnerability Notes
Latest news
Latest news
雷峰网
雷峰网
D
Darknet – Hacking Tools, Hacker News & Cyber Security
C
Cisco Blogs
H
Help Net Security
L
LINUX DO - 最新话题
L
LINUX DO - 热门话题

Show HN

GitHub - steveking-gh/firmion: Firmion is DSL and engine for firmware image generation. GitHub - villagesql/villagesql-skills: Agent skills for VillageSQL - gemini-cli-extension; claude-code-plugin GitHub - 0gsd/enough: a personal language system for planning, writing, and translation. GitHub - flightdeckhq/flightdeck: Observability and control plane for AI agents. CSP Radar GitHub - Light-Heart-Labs/DreamServer: Turn your PC, Mac, or Linux box into an AI server. LLM inference, chat UI, voice, agents, workflows, RAG, and image generation. GitHub - Diplomat-ai/diplomat-agent-ts: What can your TypeScript AI agent do to the real world? Scan your code. See which tool calls have zero checks Code Block Selector - Visual Studio Marketplace Prometheus dependency graph — interactive showcase | Riftmap Show HN: I made a vi-like modal keyboard plugin for Figma GitHub - run-llama/liteparse: A fast, helpful, and open-source document parser GitHub - dalemyers/Roar: A macOS CLI tool for notifications GitHub - district-solutions/open-agent-tools-coder: Enables small-to-large self-hosted ai models to use local source code when running tool-calling agentic workloads. We actively data mine 20,900+ (2+ TB) popular github repos using large and small ai models to create reuseable: json, markdown and parquet files for local-first tool-calling models. GitHub - progapandist/stripeek: A local TUI proxy for real-time Stripe API debugging, built for navigating complex payloads fast. GitHub - sir1st/hermes-desktop: All-in-one cross-platform desktop app for Hermes Agent — bundles Python + hermes-agent + hermes-web-ui GitHub - astefanutti/shaderbang: Shebang for Shaders Show HN: Generate Claude Code Workflows using Spec Driven Development approach GitHub - nixys/nxs-universal-chart: The Helm chart you can use to install any of your applications into Kubernetes/OpenShift Show HN: AI agents for UK GDAD PCF roles and their skills The Two Pillars: Mixer Mode and Meta-Software in the Reorganization of Software Work After AI GitHub - JaiCode08/teleport-env What 1,000+ Harness Experiments Taught Me About Self-Improving Agents Show HN: Liiists, a Markdown-first, iOS and CLI list app SwiperTab – Get this Extension for 🦊 Firefox (en-US) GitHub - kouhxp/fftext: Summarize, explain, fact-check, or translate any text, URL, or file. No GPU. No cloud. One command GitHub - sweetpad-dev/sweetpad: Develop Swift/iOS projects using VSCode GitHub - dogmaticdev/IRON: IRON a.k.a. Intermediate Representation Object Notation is a Interpreter/Database that is used to create Programming Languages. GitHub - sjhalani7/vaen: Package your AI coding harness into a portable .agent file, and share it across repos, teams, & the community without ever having to copy-paste instructions, skills, MCP config, or secrets. Show HN: Gandalf the Grader Show HN: Citadeld – replay any CI failure locally from a single file GitHub - tdortman/cuSBF: High-Performance GPU Super Bloom Filter coral-ai/claude-code-token-xray at main · Coral-Bricks-AI/coral-ai GitHub - ulyssestenn/funes: Funes is a Git-based framework for LLM-managed knowledge work: an AI Librarian ingests raw sources, builds an interlinked Markdown knowledge base, and uses it to produce cited reports, analyses, and other outputs. GitHub - ThatXliner/gah: Git Add Hunk, built for agents to use GitHub - harmont-dev/harmont-cli: Command-line client for the Harmont CI platform GitHub - brooksmcmillin/mcp-authflow: OAuth 2.0 Authorization Server framework for MCP servers GitHub - javaid-codes/audit-supply-chain-agents GitHub - amorey/gochan: A small library of common channel architectures for Go, inspired by Rust GitHub - arifozgun/OpenGem: Free, Open-Source AI API Gateway with Gemini, OpenAI & Anthropic Compatibility in 1 file GitHub - Pranesh950/BioPetals: 🌸 Run BIOxAI models at home, BitTorrent-style. Fine-tuning and inference up to 10x faster than offloading GitHub - cnguyen14/bounty-doctor: Diagnose a GitHub bounty issue before you waste hours: detects honeypot scam repos, AI-bot attempt swarms, and stale contests. Show HN: CoreMCP – MCP Server for On-Prem DBs Show HN: KittyHTML – Render HTML/CSS as an inline image in your terminal GitHub - bingud/filemat: Web-based file manager Show HN: TruthLens – Free multi-signal deepfake image detector GitHub - apexlocal-jz/claude-usage-tray: Windows system-tray app showing your Claude Code rate-limit usage at a glance. Zero deps, ~300 lines of PowerShell. Cross-IDE (works regardless of VS Code, Cursor, plain terminal). Release v0.1.2.1 · kouhxp/yapsnap GitHub - noopolis/moltnet: Self-hostable chat network for AI agents. Pre-built bridges for Claude Code, Codex, and the Claws. Rooms, DMs, history. No Slack bots, no Matrix, no glue code. GitHub - tamerh/enju: Coordinating Humans, AI Agents, and Compute as Peers on a Shared Workflow Graph Show HN: Continuity-auth – Respect-weighted rate limits for the open web GitHub - luml-ai/luml: AI lifecycle platform where engineers and agents track experiments, train models, and ship to production. GitHub - mrdanielcasper/CoreTex: A UNIX-inspired, biomimetic, flat-file AI harness and knowledge engine. GitHub - clemg/pierre-github: Pierre's diffs.com and trees.software for Github GitHub - lyriks-io/unspaghettit: Behavior-driven AI development without prompt spaghetti. GitHub - dotexorg/saferpc: Typed, end-to-end encrypted RPC over any bidirectional channel. GitHub - BeeZeeAgent/beezee: Agent harness orchestration Legato Next.js Boilerplate for Internal Tools · CoreUI GitHub - clark-labs-inc/clark-hash: Clark Hash, 32x smaller searchable sketches for embeddings GitHub - ZeroPointRepo/youtube-mcp: The fastest YouTube transcript + YouTube search MCP for AI agents. Try for free. Typing Mastery — climb toward 100+ WPM, deliberately GitHub - Andebugulin/Awareen GitHub - fayzan123/claude-workflow-composer: Visual desktop app for composing multi-agent coding workflows. Drag agents, attach skills and MCPs, wire handoffs, export to .claude/ GitHub - harshaneel/humanize: Best static AI text humanizer. Two research-grounded skills that work in any LLM (Claude, ChatGPT, Gemini, Codex): humanize beats perplexity-based detectors, ai-check produces forensic scoring with evidence-quoted flags. Nine levers, 50+ peer-reviewed sources, 2024-2026 detection literature. GitHub - StackOneHQ/stack-nudge GitHub - nodes-app/swift-markdown-engine: A native AppKit Markdown editor for macOS, built on TextKit 2 and bridged to SwiftUI. We hardened an LLM agent. Each defense we added made it more exploitable. GitHub - alkait/WhatsKept: Agent-queryable WhatsApp history from an iOS backup — a single Go binary. GitHub - octelium/cordium: Open-source, general-purpose sandbox platform for devs and AI agents that provides identity-based secure access to infrastructure without credentials. WAR.GOV/UFO Microfilm5 GitHub - scosman/videowright: Build animated explainer videos with your coding agent GitHub - dipankar/dscode: The code editor you can take apart. GitHub - zoharbabin/web-researcher-mcp: MCP server (Go) for AI assistants: web search, content extraction, academic/patent/news research. Multi-provider routing, 4-tier scraping, search lenses. Works with Claude, Cursor, and any MCP client. GitHub - ruvnet/RuView: π RuView turns commodity WiFi signals into real-time spatial intelligence, vital sign monitoring, and presence detection — all without a single pixel of video. GitHub - scanaislop/aislop: Catch the slop AI coding agents leave in your code: narrative comments, swallowed exceptions, as-any casts, dead code, oversized functions. 50+ rules across 7 languages (TypeScript, JavaScript, Python, Go, Rust, Ruby, PHP). Sub-second, deterministic, no LLM at runtime. MIT-licensed. GitHub - kouhxp/cheap-im: CPU-only voice agent approximating Thinking Machines' Interaction Models demo GitHub - unprovable/OrchidMantis: Orchid Mantis — standalone framework for Zero-Knowledge Proofs of eXploit (ZKPoX). GitHub - MarcellM01/TinySearch: Shrink the web for your local LLMs! GitHub - TangibleResearch/Halgorithem: A Algo designed to detect AI Hallucitions GitHub - DO-SAY-GO/freelang: I love freelang GitHub - CarpseDeam/Aura-IDE: An AI coding harness that shaped itself - Planner/Worker agents, repo awareness, surgical edits, validation, recovery, and safe diff approvals. GitHub - chojs23/concord: A feature-rich TUI client for Discord GitHub - tommyjepsen/awesome-ux-skills: UX & AI Product designs skills you can use today in Claude Code GitHub - aerf-spec/aerf: Agent Evidence Receipt Format (AERF) — an open specification for tamper-evident, independently verifiable records of AI agent actions. GitHub - kklimuk/docx-cli: CLI for AI agents (Claude, Codex) to read, edit, and comment on .docx files with full format fidelity. GitHub - Jwrede/tokentoll: Catch LLM cost changes in code review. Infracost for LLM spend. GitHub - samchon/ttsc: A `typescript-go` toolchain for compiler-powered plugins and type-safe execution + 500x faster lint integrated into compiler GitHub - Higangssh/homebutler: 🏠 Manage your homelab from chat. Single binary, zero dependencies. GitHub - olalie/tapmap: See where your computer connects and what stands out on a live world map. GitHub - Diplomat-ai/diplomat-agent: What can your AI agent do to the real world? Scan your code. See which tool calls have zero checks GitHub - Bajusz15/beacon: Open-source agent for secure remote access, monitoring, and deploys across home-lab and self-hosted machines like Raspberry Pi, N100, or any Linux server. Open web based TTY or tunnel Home Assistant and other local services securely without opening ports. BigTech AI News - Chrome 应用商店 GitHub - vinhnx/VTCode: VT Code is an open-source coding agent with LLM-native code understanding and robust shell safety. Supports multiple LLM providers with automatic failover and efficient context management. GitHub - michaelaz774/decision-engine: A decision operating system for startup founders, powered by Claude Code. Synthesizes wisdom from 25+ legendary founders and investors into interactive AI-driven decision frameworks. GitHub - Chrilleweb/dotenv-diff: Validate environment variable usage in your codebase GitHub - Lumen-Labs/brainapi2: BrainAPI is a knowledge graph–powered AI memory layer that transforms unstructured data into structured knowledge, enabling intelligent search, recommendations, and contextual memory for AI agents and applications. GitHub - familiar-software/familiar: Let AI watch you work. Familiar lets your AI update its memory, skills, and knowledge by watching your screen. GitHub - skorotkiewicz/rudo: A small, elegant dock for Wayland GitHub - muxshed/shed: One stream in, or many. Every destination, simultaneously. No cloud middleman, no per-channel fees, no limits. make sidebar/address bar rounded corner toggleable
GitHub - sofumel/claude-handoff-revive: Resume Claude Code work after rate/usage/context limits without replaying the prior transcript. Auto-saves at 90%/95% usage. Plugin-installable, 10 languages.
sofumel · 2026-05-27 · via Show HN

claude-handoff-revive

CI Release Stars Contributors License: MIT Multi-language

Continue Claude Code work after a rate / usage / context limit without replaying your prior session — typically tens of thousands of tokens, often 100k+ for long sessions.

🇺🇸 English · 🇯🇵 日本語 · 🇨🇳 中文 · 🇹🇼 繁體中文 · 🇰🇷 한국어 · 🇪🇸 Español · 🇵🇹 Português · 🇩🇪 Deutsch · 🇫🇷 Français · 🇹🇷 Türkçe


🇺🇸 English

The problem

When Claude Code shows You've hit your limit · resets ..., the standard recovery — claude --resume or claude -c — reloads the entire prior conversation into context. A medium session typically burns tens of thousands of tokens, often 100k+ for long sessions, before you've asked a single question.

What this skill does

It saves only the minimum needed to continue, structured into .claude/handoff/current.md (~1–3k tokens):

  • goal — what you were trying to accomplish
  • done / wip / todo — completed, in-progress, not-yet-started tasks
  • next_action — the concrete next step (file:line + exact command)
  • touched_files — what files were touched and why
  • decisions — design choices, with the reason behind each

To resume, start a new session (don't use --resume). The skill reads only that one file and picks up where you left off.

Install

Plugin install (recommended) — run these inside Claude Code:

/plugin marketplace add sofumel/claude-handoff-revive       # 1. register marketplace
/plugin install handoff-revive@handoff-revive-marketplace   # 2. install plugin

Skill, slash commands, and hooks auto-activate. No settings.json editing required.

/plugin isn't available in this environment? Your Claude Code version is too old. Check with claude --version and update via brew upgrade claude-code (Homebrew) or npm update -g @anthropic-ai/claude-code (npm). If you can't update, use the manual install below.

Manual install (always works, no plugin support required):

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

# Or globally for all projects:
./install.sh --global       # .\install.ps1 -Global

For the manual install, optionally enable hooks by merging the snippet from HOOK_SETUP.md into your .claude/settings.json (the plugin install does this automatically).

How to use

Quick reference:

Action Type this in Claude Code
Save handoff /handoff-revive:save
Save with extra quality check /handoff-revive:save-strict
Resume in a NEW session /handoff-revive:resume
Toggle auto-save for this session /handoff-revive:auto on / off / status

Step 1 — when you're approaching a rate limit, save

In Claude Code, type the following command:

/handoff-revive:save

Behind the scenes, the skill runs these steps automatically (all zero-LLM-token where possible):

  1. Detects your language (10 supported: en / ja / zh / zh-TW / ko / es / pt / de / fr / tr) from your message
  2. Pulls changed files from git status — you don't have to remember which files you edited
  3. Writes a schema-conformant handoff to .claude/handoff/current.md
  4. Validates the schema, cleans up dead paths, strips empty sections — pure shell, no tokens
  5. Displays the estimated savings in the chat

Step 2 — wait for the rate limit to reset

When the time on resets ... passes, your usage window opens again.

Step 3 — start a NEW session (do NOT use --resume)

claude

When the new session starts:

  • The SessionStart hook detects .claude/handoff/current.md (if modified within the last 7 days)
  • It injects context for Claude: "A recent handoff exists — prefer it over --resume"
  • You run /handoff-revive:resume and Claude reads only the handoff file and continues

Step 4 — (optional) higher-quality saves

/handoff-revive:save-strict

After the regular save, Claude runs an extra semantic-quality check:

  • Is goal specific enough that someone with no context understands it?
  • Does each decision include a real reason, not just the decision restated?
  • Are blockers phrased as answerable questions?
  • Does wip describe state precisely enough to resume mid-edit?

Vague fields get rewritten automatically.

Optional: turn-count nudge

If the Stop hook is enabled, Claude reminds you to checkpoint every N turns (default 15):

[handoff-revive] Turn 15 — checkpoint due. Run /handoff-revive:save to save.

Plugin install activates this automatically; manual install — see HOOK_SETUP.md. Override the threshold with HANDOFF_CHECKPOINT_EVERY=10.

Auto-save at 90% / 95% usage (synced with the UI notification)

The usage-monitor hook (PostToolUse) reads rate_limits.five_hour.used_percentage — the same value that drives Claude Code's "approaching usage limit" notification. When you cross 90%, Claude auto-saves a full handoff before its next response (no asking, no interruption). At 95%, it saves again with an urgent notice.

Per-session opt-out if you don't want auto-save in the current thread:

/handoff-revive:auto off       # disable for THIS session
/handoff-revive:auto on        # re-enable
/handoff-revive:auto status    # show current state + thresholds

New Claude sessions reset to enabled. To disable globally: export HANDOFF_AUTO_SAVE_PERCENT=disabled HANDOFF_URGENT_PERCENT=disabled.

To shift thresholds: export HANDOFF_AUTO_SAVE_PERCENT=80 (earlier), or set per-hook in settings.json (see HOOK_SETUP.md).

Token budget

Method Tokens replayed to resume
claude --resume tens of thousands, often 100k+
claude -c tens of thousands, often 100k+
handoff-revive 1,000–3,000

You break even after just one save and one resume. Even in the worst case where you save and never resume, the overhead is only a few KB — costs never run away.

License

MIT — see LICENSE.


🇯🇵 日本語

解決する問題

You've hit your limit · resets ... という制限通知が表示された後、claude --resume-c で再開すると、それまでの会話履歴がまるごとコンテキストに再ロードされます。中規模のセッションでも数万トークン、長いセッションでは 10 万を超えるトークン分が、まだ何も質問していない段階で消費されてしまいます。

この skill が提供するもの

作業の続きに必要な最小限の情報だけを構造化して .claude/handoff/current.md に保存します(約 1〜3k トークン)。保存される項目:

  • goal — 何を達成しようとしていたか
  • done / wip / todo — 完了済み / 進行中 / 未着手のタスク
  • next_action — 実行可能な次の一手(file:line + 具体的なコマンド)
  • touched_files — 触ったファイルとその理由
  • decisions — 設計判断と、その根拠

再開時は claude --resume を使わず 新規セッションを起動するだけ。skill がこの 1 ファイルだけを読み込み、即座に作業を引き継ぎます。

インストール

プラグインインストール (推奨) — Claude Code 内で実行:

/plugin marketplace add sofumel/claude-handoff-revive       # 1. マーケットプレイス登録
/plugin install handoff-revive@handoff-revive-marketplace   # 2. プラグインインストール

Skill・スラッシュコマンド・hook がすべて自動で有効化されます。settings.json の編集は不要。

/plugin isn't available in this environment と表示された場合、Claude Code のバージョンが古い可能性があります。claude --version で確認し、brew upgrade claude-code (Homebrew) または npm update -g @anthropic-ai/claude-code (npm) で更新してください。それでも使えない場合は下のマニュアルインストールを使用してください。

マニュアルインストール (どの環境でも動作):

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

# 全プロジェクト共通でインストール:
./install.sh --global       # .\install.ps1 -Global

マニュアルインストールの場合のみ、HOOK_SETUP.md のスニペットを .claude/settings.json にマージして hooks を有効化してください (プラグインインストールでは自動)。

使い方

早見表:

やりたいこと Claude Code に入力するコマンド
ハンドオフを保存 /handoff-revive:save
品質チェック付きで保存 /handoff-revive:save-strict
新規セッション で再開 /handoff-revive:resume
このセッションの自動保存を切り替え /handoff-revive:auto on / off / status

ステップ 1 — レート制限が近づいたら保存する

Claude Code で、以下のコマンドを入力してください:

/handoff-revive:save

裏では skill が以下を自動実行します(できる限りトークンを使わずに):

  1. 言語を検出 (10 言語対応: en / ja / zh / zh-TW / ko / es / pt / de / fr / tr) — あなたのメッセージから判定
  2. git status から変更ファイルを抽出 — どのファイルを編集したか思い出す必要なし
  3. スキーマ通りのハンドオフを .claude/handoff/current.md に書き込み
  4. スキーマ検証 + 無効パスエントリの自動削除 + 空セクションの除去 — すべて純シェル (Claude を呼ばずシェルスクリプトだけで処理) なので LLM トークンを消費しません
  5. 節約見込みをチャットに表示

ステップ 2 — レート制限の解除を待つ

resets ... の時刻を過ぎたら、使用枠が回復します。

ステップ 3 — 新規セッションを起動 (--resume は使わない)

claude

新セッションが起動すると:

  • SessionStart hook が .claude/handoff/current.md (7 日以内に更新) を検出
  • Claude にコンテキスト注入: 「最近のハンドオフがあります。--resume ではなくこちらを優先してください」
  • あなたが /handoff-revive:resume を実行すると、Claude がハンドオフファイルだけを読んで作業再開

ステップ 4 — (任意) 保存品質を強化したい場合

/handoff-revive:save-strict

通常の保存後に、Claude による追加の意味的チェックが走ります:

  • goal が具体的かどうか (「認証バグを直す」のような曖昧さでなく「login.ts の bcrypt.compare をタイミング攻撃に強い実装に置換」のような具体性)
  • decision根拠が書いてあるか (決定の言い換えではなく "WHY" が明示されているか)
  • blockers が「答えられる質問形式」になっているか
  • wip が「再開時に作業位置が分かる粒度」になっているか

不十分なフィールドを Claude が自動的に書き直します。

任意機能: ターン数 nudge

Stop hook を有効化すると、ターン数が指定値(デフォルト 15)を超えるたびにリマインドが出ます:

[handoff-revive] Turn 15 — checkpoint due. Run /handoff-revive:save to save.

プラグインインストールでは自動有効化されます。マニュアルインストールの場合は HOOK_SETUP.md を参照。閾値は環境変数 HANDOFF_CHECKPOINT_EVERY=10 で変更可能。

90% / 95% 到達で自動保存 (UI 通知と同期)

usage-monitor hook (PostToolUse) が rate_limits.five_hour.used_percentage を読み取ります — Claude Code の「使用量の上限に近づいています」通知を駆動するのと同じ値です。90% 到達時、Claude が次の応答前にフルクオリティのハンドオフを自動保存します (確認なし、作業中断なし)。95% に達すると、緊急通知付きで再度自動保存されます。

このセッションの自動保存をオフにしたい場合:

/handoff-revive:auto off       # このセッションのみ無効化
/handoff-revive:auto on        # 再有効化
/handoff-revive:auto status    # 現在の状態 + 閾値を表示

新しい Claude セッションを起動すると自動保存はデフォルト ON に戻ります。グローバルに無効化するには: export HANDOFF_AUTO_SAVE_PERCENT=disabled HANDOFF_URGENT_PERCENT=disabled

閾値変更: export HANDOFF_AUTO_SAVE_PERCENT=80 (より早めに発火)、または settings.json の hook で個別指定 (HOOK_SETUP.md 参照)。

トークン収支

方式 再開時に再ロードされるトークン
claude --resume 数万、長セッションでは 10 万超も
claude -c 数万、長セッションでは 10 万超も
handoff-revive 1,000〜3,000

1 回保存して 1 回再開すれば、必ず黒字です。 たとえ保存だけで一度も再開しない場合でも、発生するのは数 KB 程度のオーバーヘッドだけで、コストが暴走することはありません。

ライセンス

MIT — LICENSE 参照。


🇨🇳 中文

要解决的问题

当 Claude Code 显示 You've hit your limit · resets ... 时,使用 claude --resumeclaude -c 恢复会把之前的整个会话重新载入上下文。中等规模会话通常需要数万 tokens,长会话经常超过 10 万——还没问一个问题就已经用掉了。

此 skill 提供什么

把恢复所需的最少信息以结构化方式保存到 .claude/handoff/current.md(约 1–3k tokens):

  • goal — 你想完成什么
  • done / wip / todo — 完成 / 进行中 / 待办
  • next_action — 可执行的下一步(file:line + 具体命令)
  • touched_files — 涉及的文件及其原因
  • decisions — 设计决策及其理由

恢复时启动新会话(不要用 --resume)。skill 只读取这一个文件即可继续工作。

安装

插件安装 (推荐) — 在 Claude Code 内运行:

/plugin marketplace add sofumel/claude-handoff-revive       # 1. 注册插件市场
/plugin install handoff-revive@handoff-revive-marketplace   # 2. 安装插件

Skill、斜杠命令、hooks 全部自动激活,无需编辑 settings.json

如果出现 /plugin isn't available in this environment,说明 Claude Code 版本过旧。运行 claude --version 确认,并通过 brew upgrade claude-codenpm update -g @anthropic-ai/claude-code 升级。

手动安装 (始终可用):

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

手动安装时可选:将 HOOK_SETUP.md 中的代码片段合并到 .claude/settings.json 启用 hooks (插件安装会自动完成)。

使用方法

快速参考:

操作 在 Claude Code 输入
保存 handoff /handoff-revive:save
带质量检查的保存 /handoff-revive:save-strict
新会话中恢复 /handoff-revive:resume
切换本会话的自动保存 /handoff-revive:auto on / off / status

步骤 1 — 接近速率限制时保存

/handoff-revive:save

skill 自动执行(尽量零 token):

  1. 检测语言 (10 种语言: en / ja / zh / zh-TW / ko / es / pt / de / fr / tr)
  2. git status 提取变更文件 — 不用回忆改了哪些文件
  3. 按 schema 写入 .claude/handoff/current.md
  4. schema 验证 + 死路径删除 + 空段落删除 — 纯 shell,零 token
  5. 在聊天中显示节省估算

步骤 2 — 等待限制重置

到达 resets ... 时间后,使用额度恢复。

步骤 3 — 启动新会话 (不要用 --resume)

claude

新会话启动时:

  • SessionStart hook 检测 .claude/handoff/current.md (7 天以内)
  • 注入上下文: "存在最近的 handoff,优先使用而不是 --resume"
  • 你运行 /handoff-revive:resume — Claude 只读 handoff 文件即可恢复

步骤 4 — (可选) 提高保存质量

/handoff-revive:save-strict

常规保存后再追加 Claude 的语义检查:

  • goal 是否具体?
  • 各 decision 是否包含真实理由?
  • blockers 是否为可回答的问题?
  • wip 是否精确到能从中间恢复?

90% / 95% 自动保存 (与 UI 通知同步)

usage-monitor hook 读取 rate_limits.five_hour.used_percentage —— 这与 Claude Code "接近使用上限" 通知所用的值相同。90% 时 Claude 在下次回复前自动完整保存 handoff (无需询问、不打断)。95% 时再次保存并附紧急通知。

本会话不需要自动保存时:

/handoff-revive:auto off       # 仅本会话禁用
/handoff-revive:auto on        # 重新启用
/handoff-revive:auto status    # 查看状态 + 阈值

新会话默认启用。全局禁用: export HANDOFF_AUTO_SAVE_PERCENT=disabled HANDOFF_URGENT_PERCENT=disabled

License

MIT — see LICENSE.


🇹🇼 繁體中文

要解決的問題

當 Claude Code 顯示 You've hit your limit · resets ... 時,使用 claude --resumeclaude -c 恢復會把之前的整個會話重新載入上下文。中等規模對話通常需要數萬 tokens,長對話經常超過 10 萬——還沒問一個問題就已經用掉了。

此 skill 提供什麼

把恢復所需的最少資訊以結構化方式儲存到 .claude/handoff/current.md(約 1–3k tokens):

  • goal — 你想完成什麼
  • done / wip / todo — 完成 / 進行中 / 待辦
  • next_action — 可執行的下一步(file:line + 具體命令)
  • touched_files — 涉及的檔案及其原因
  • decisions — 設計決策及其理由

恢復時啟動新會話(不要用 --resume)。skill 只讀取這一個檔案即可繼續工作。

安裝

外掛程式安裝 (推薦) — 在 Claude Code 內執行:

/plugin marketplace add sofumel/claude-handoff-revive       # 1. 註冊 marketplace
/plugin install handoff-revive@handoff-revive-marketplace   # 2. 安裝外掛

如果出現 /plugin isn't available in this environment,說明 Claude Code 版本過舊。執行 claude --version 確認,並透過 brew upgrade claude-codenpm update -g @anthropic-ai/claude-code 升級。

手動安裝 (始終可用):

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

使用方式

操作 在 Claude Code 輸入
儲存 handoff /handoff-revive:save
帶品質檢查的儲存 /handoff-revive:save-strict
新會話中恢復 /handoff-revive:resume
切換本會話的自動儲存 /handoff-revive:auto on / off / status

License

MIT — see LICENSE.


🇰🇷 한국어

해결하는 문제

You've hit your limit · resets ... 메시지가 나타난 뒤 claude --resume 또는 -c 로 재개하면 이전 세션 전체 가 컨텍스트로 다시 로드됩니다. 중간 규모 세션이라면 수만 토큰, 긴 세션에서는 10만+ 토큰이 아무 질문하기도 전에 사라집니다.

이 skill 이 제공하는 것

작업 재개에 필요한 최소 정보만을 구조화하여 .claude/handoff/current.md 에 저장합니다(약 1–3k tokens):

  • goal — 무엇을 달성하려 했는가
  • done / wip / todo — 완료 / 진행 중 / 대기
  • next_action — 실행 가능한 다음 단계 (file:line + 구체적 명령)
  • touched_files — 다룬 파일과 그 이유
  • decisions — 설계 결정과 근거

재개할 때는 --resume 을 쓰지 말고 새 세션을 시작합니다. skill 이 그 파일 하나만 읽고 즉시 작업을 이어갑니다.

설치

플러그인 설치 (권장) — Claude Code 안에서 실행:

/plugin marketplace add sofumel/claude-handoff-revive       # 1. 마켓플레이스 등록
/plugin install handoff-revive@handoff-revive-marketplace   # 2. 플러그인 설치

/plugin isn't available in this environment 메시지가 나오면 Claude Code 버전이 오래되었습니다. claude --version 으로 확인하고 brew upgrade claude-code 또는 npm update -g @anthropic-ai/claude-code 로 업데이트하세요.

수동 설치 (항상 동작):

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

사용법

빠른 참조:

작업 Claude Code 입력
핸드오프 저장 /handoff-revive:save
품질 검사 포함 저장 /handoff-revive:save-strict
새 세션 에서 재개 /handoff-revive:resume
이 세션 자동 저장 토글 /handoff-revive:auto on / off / status

단계 1 — 사용량 한도가 가까워지면 저장

/handoff-revive:save

skill 이 자동으로 (가능한 한 0 토큰으로):

  1. 언어 감지 (10개 언어 지원: en / ja / zh / zh-TW / ko / es / pt / de / fr / tr)
  2. git status 에서 변경 파일 추출
  3. 스키마대로 .claude/handoff/current.md 에 저장
  4. 스키마 검증 + 죽은 경로 제거 + 빈 섹션 제거 — 순 shell, 0 토큰
  5. 절약량 표시

단계 2 — 한도 리셋 대기

resets ... 시간이 지나면 사용량이 복구됩니다.

단계 3 — 새 세션 시작 (--resume 사용 금지)

claude

새 세션 시작 시:

  • SessionStart hook 이 .claude/handoff/current.md 감지 (7일 이내)
  • 컨텍스트 주입: "최근 핸드오프 존재 — --resume 대신 이걸 사용"
  • /handoff-revive:resume 를 실행하면 Claude 가 핸드오프만 읽고 재개

단계 4 — (선택) 더 높은 품질의 저장

/handoff-revive:save-strict

추가로 Claude 의 의미적 검증 실행: goal 이 구체적인지, 각 decision 에 진짜 이유가 있는지 등.

90% / 95% 자동 저장 (UI 알림과 동기)

usage-monitor hook 이 rate_limits.five_hour.used_percentage 를 읽음 — Claude Code 의 "사용량 한도 접근" 알림과 같은 값. 90% 도달 시 Claude 가 다음 응답 전에 풀 핸드오프를 자동 저장 (질문 없이, 중단 없이). 95% 에서 긴급 알림과 함께 재저장.

이 세션에서 자동 저장 끄기:

/handoff-revive:auto off       # 이 세션만 비활성화
/handoff-revive:auto on        # 다시 활성화
/handoff-revive:auto status    # 상태 + 임계값 확인

새 세션은 기본 활성화. 전역 비활성화: export HANDOFF_AUTO_SAVE_PERCENT=disabled HANDOFF_URGENT_PERCENT=disabled.

License

MIT — see LICENSE.


🇪🇸 Español

El problema

Cuando Claude Code muestra You've hit your limit · resets ..., la recuperación habitual — claude --resume o -c — recarga toda la conversación anterior en el contexto. Una sesión mediana consume decenas de miles de tokens, a menudo 100k+ en sesiones largas, antes de hacer una sola pregunta.

Lo que ofrece este skill

Guarda solo el mínimo necesario para continuar, estructurado en .claude/handoff/current.md (~1–3k tokens):

  • goal — qué intentabas lograr
  • done / wip / todo — completado / en progreso / pendiente
  • next_action — siguiente paso ejecutable (file:line + comando exacto)
  • touched_files — archivos tocados y su razón
  • decisions — decisiones de diseño y el por qué

Para reanudar, inicia una sesión nueva (NO uses --resume). El skill lee solo ese archivo y continúa.

Instalación

Plugin (recomendado) — ejecutar dentro de Claude Code:

/plugin marketplace add sofumel/claude-handoff-revive       # 1. registrar marketplace
/plugin install handoff-revive@handoff-revive-marketplace   # 2. instalar plugin

Si aparece /plugin isn't available in this environment, tu versión de Claude Code es antigua. Verifica con claude --version y actualiza con brew upgrade claude-code o npm update -g @anthropic-ai/claude-code.

Instalación manual (siempre funciona):

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

Cómo usar

Referencia rápida:

Acción Escribe en Claude Code
Guardar handoff /handoff-revive:save
Guardar con verificación /handoff-revive:save-strict
Reanudar en una sesión nueva /handoff-revive:resume
Alternar auto-guardado en esta sesión /handoff-revive:auto on / off / status

Paso 1 — al acercarte al límite, guarda

/handoff-revive:save

El skill ejecuta automáticamente (con cero tokens cuando es posible):

  1. Detecta tu idioma (10 idiomas: en / ja / zh / zh-TW / ko / es / pt / de / fr / tr)
  2. Extrae archivos modificados de git status
  3. Escribe el handoff conforme al schema en .claude/handoff/current.md
  4. Valida schema + elimina rutas muertas + elimina secciones vacías — solo shell, cero tokens
  5. Muestra el ahorro estimado

Paso 2 — espera al reset

Cuando pase la hora indicada en resets ..., tu cuota se restablece.

Paso 3 — inicia una sesión nueva (NO uses --resume)

claude

Al iniciar:

  • El hook SessionStart detecta .claude/handoff/current.md (modificado en últimos 7 días)
  • Inyecta contexto: "Existe handoff reciente — prefiérelo sobre --resume"
  • Ejecuta /handoff-revive:resume — Claude lee solo el handoff y continúa

Paso 4 — (opcional) guardado de mayor calidad

/handoff-revive:save-strict

Tras el guardado normal, Claude ejecuta una verificación semántica: ¿es goal específico? ¿incluye cada decision una razón real?

Auto-guardado al 90% / 95% (sincronizado con la notificación de UI)

El hook usage-monitor lee rate_limits.five_hour.used_percentage — el mismo valor que dispara la notificación "uso cerca del límite" de Claude Code. Al 90% Claude auto-guarda un handoff completo antes de su siguiente respuesta (sin preguntar, sin interrumpir). Al 95% vuelve a guardar con aviso urgente.

Para desactivar el auto-guardado en esta sesión:

/handoff-revive:auto off       # desactivar SOLO esta sesión
/handoff-revive:auto on        # reactivar
/handoff-revive:auto status    # ver estado + umbrales

Las sesiones nuevas vuelven al estado activado. Desactivación global: export HANDOFF_AUTO_SAVE_PERCENT=disabled HANDOFF_URGENT_PERCENT=disabled.

License

MIT — see LICENSE.


🇵🇹 Português

O problema

Quando o Claude Code mostra You've hit your limit · resets ..., a recuperação habitual — claude --resume ou -c — recarrega toda a conversa anterior no contexto. Uma sessão média gasta dezenas de milhares de tokens, frequentemente 100k+ em sessões longas, antes de fazer uma única pergunta.

O que esta skill faz

Guarda apenas o mínimo necessário para continuar, estruturado em .claude/handoff/current.md (~1–3k tokens):

  • goal — o que estavas a tentar alcançar
  • done / wip / todo — concluído / em curso / pendente
  • next_action — próximo passo executável (file:line + comando exato)
  • touched_files — ficheiros tocados e a razão
  • decisions — decisões de design e o porquê

Para retomar, inicia uma sessão nova (NÃO uses --resume). A skill lê apenas esse ficheiro e continua.

Instalação

Plugin (recomendado) — executar dentro do Claude Code:

/plugin marketplace add sofumel/claude-handoff-revive       # 1. registar marketplace
/plugin install handoff-revive@handoff-revive-marketplace   # 2. instalar plugin

Se aparecer /plugin isn't available in this environment, a tua versão do Claude Code é antiga. Verifica com claude --version e atualiza via brew upgrade claude-code ou npm update -g @anthropic-ai/claude-code.

Instalação manual (sempre funciona):

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

Uso

Ação Escreve no Claude Code
Guardar handoff /handoff-revive:save
Guardar com verificação /handoff-revive:save-strict
Retomar numa sessão nova /handoff-revive:resume
Alternar auto-save nesta sessão /handoff-revive:auto on / off / status

License

MIT — see LICENSE.


🇩🇪 Deutsch

Das Problem

Wenn Claude Code You've hit your limit · resets ... anzeigt, lädt die übliche Wiederherstellung — claude --resume oder -cdie gesamte vorherige Konversation wieder in den Kontext. Eine mittelgroße Sitzung verbraucht zehntausende Tokens, oft 100k+ bei langen Sitzungen, bevor du eine einzige Frage gestellt hast.

Was dieser Skill tut

Speichert nur das Minimum, das zum Fortsetzen nötig ist, strukturiert in .claude/handoff/current.md (~1–3k Tokens):

  • goal — was du erreichen wolltest
  • done / wip / todo — erledigt / in Bearbeitung / ausstehend
  • next_action — konkreter nächster Schritt (file:line + exaktes Kommando)
  • touched_files — bearbeitete Dateien und der Grund
  • decisions — Design-Entscheidungen und das Warum

Zum Fortsetzen startest du eine neue Sitzung (KEIN --resume). Der Skill liest nur diese eine Datei und setzt fort.

Installation

Plugin (empfohlen) — in Claude Code ausführen:

/plugin marketplace add sofumel/claude-handoff-revive       # 1. Marketplace registrieren
/plugin install handoff-revive@handoff-revive-marketplace   # 2. Plugin installieren

Falls /plugin isn't available in this environment erscheint, ist deine Claude Code Version zu alt. Prüfe mit claude --version und aktualisiere via brew upgrade claude-code oder npm update -g @anthropic-ai/claude-code.

Manuelle Installation (funktioniert immer):

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

Benutzung

Aktion In Claude Code eingeben
Handoff speichern /handoff-revive:save
Mit Qualitätsprüfung speichern /handoff-revive:save-strict
In neuer Sitzung fortsetzen /handoff-revive:resume
Auto-Save für diese Sitzung umschalten /handoff-revive:auto on / off / status

License

MIT — see LICENSE.


🇫🇷 Français

Le problème

Quand Claude Code affiche You've hit your limit · resets ..., la récupération habituelle — claude --resume ou -c — recharge toute la conversation précédente dans le contexte. Une session de taille moyenne consomme des dizaines de milliers de tokens, souvent 100k+ pour les sessions longues, avant que vous n'ayez posé une seule question.

Ce que fait ce skill

Sauvegarde uniquement le minimum nécessaire pour continuer, structuré dans .claude/handoff/current.md (~1–3k tokens) :

  • goal — ce que vous essayiez d'accomplir
  • done / wip / todo — terminé / en cours / à faire
  • next_action — prochaine étape exécutable (file:line + commande exacte)
  • touched_files — fichiers modifiés et la raison
  • decisions — décisions de conception et le pourquoi

Pour reprendre, lancez une nouvelle session (PAS --resume). Le skill lit uniquement ce fichier et continue.

Installation

Plugin (recommandé) — exécuter dans Claude Code :

/plugin marketplace add sofumel/claude-handoff-revive       # 1. enregistrer le marketplace
/plugin install handoff-revive@handoff-revive-marketplace   # 2. installer le plugin

Si /plugin isn't available in this environment apparaît, votre version de Claude Code est trop ancienne. Vérifiez avec claude --version et mettez à jour via brew upgrade claude-code ou npm update -g @anthropic-ai/claude-code.

Installation manuelle (fonctionne toujours) :

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

Utilisation

Action Tapez dans Claude Code
Sauvegarder le handoff /handoff-revive:save
Sauvegarder avec vérification qualité /handoff-revive:save-strict
Reprendre dans une NOUVELLE session /handoff-revive:resume
Basculer l'auto-sauvegarde pour cette session /handoff-revive:auto on / off / status

License

MIT — see LICENSE.


🇹🇷 Türkçe

Sorun

Claude Code You've hit your limit · resets ... mesajını gösterdiğinde, olağan kurtarma — claude --resume veya -ctüm önceki konuşmayı bağlama yeniden yükler. Orta ölçekli bir oturum, tek bir soru sormadan önce on binlerce token, uzun oturumlarda sıklıkla 100k+ tüketir.

Bu skill ne yapar

Devam etmek için yalnızca gerekli olan minimumu yapılandırılmış olarak .claude/handoff/current.md dosyasına kaydeder (~1–3k token):

  • goal — neyi başarmaya çalıştığınız
  • done / wip / todo — tamamlandı / devam ediyor / beklemede
  • next_action — somut sonraki adım (file:line + tam komut)
  • touched_files — dokunulan dosyalar ve nedeni
  • decisions — tasarım kararları ve nedenleri

Devam etmek için yeni bir oturum başlatın (--resume KULLANMAYIN). Skill yalnızca bu dosyayı okur ve devam eder.

Kurulum

Plugin (önerilen) — Claude Code içinde çalıştırın:

/plugin marketplace add sofumel/claude-handoff-revive       # 1. marketplace'i kaydet
/plugin install handoff-revive@handoff-revive-marketplace   # 2. plugin'i yükle

/plugin isn't available in this environment görürseniz, Claude Code sürümünüz çok eski. claude --version ile kontrol edin ve brew upgrade claude-code veya npm update -g @anthropic-ai/claude-code ile güncelleyin.

Manuel kurulum (her zaman çalışır):

git clone https://github.com/sofumel/claude-handoff-revive.git
cd claude-handoff-revive
./install.sh /path/to/your-project        # Linux/macOS/WSL/Git-Bash
# .\install.ps1 -Target C:\path\to\proj   # Windows PowerShell

Kullanım

Eylem Claude Code'a yazın
Handoff kaydet /handoff-revive:save
Kalite kontrolüyle kaydet /handoff-revive:save-strict
YENİ oturumda devam et /handoff-revive:resume
Bu oturum için otomatik kaydı değiştir /handoff-revive:auto on / off / status

License

MIT — see LICENSE.