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

S
Schneier on Security
cs.AI updates on arXiv.org
cs.AI updates on arXiv.org
T
Threat Research - Cisco Blogs
C
Cyber Attacks, Cyber Crime and Cyber Security
C
CXSECURITY Database RSS Feed - CXSecurity.com
A
Arctic Wolf
Security Latest
Security Latest
Simon Willison's Weblog
Simon Willison's Weblog
I
Intezer
cs.CV updates on arXiv.org
cs.CV updates on arXiv.org
T
Troy Hunt's Blog
Latest news
Latest news
Help Net Security
Help Net Security
S
Security Affairs
Webroot Blog
Webroot Blog
The Hacker News
The Hacker News
AI
AI
cs.CL updates on arXiv.org
cs.CL updates on arXiv.org
T
Tor Project blog
Forbes - Security
Forbes - Security
Google DeepMind News
Google DeepMind News
AWS News Blog
AWS News Blog
Attack and Defense Labs
Attack and Defense Labs
P
Proofpoint News Feed
www.infosecurity-magazine.com
www.infosecurity-magazine.com
H
Help Net Security
L
Lohrmann on Cybersecurity
S
SegmentFault 最新的问题
Google Online Security Blog
Google Online Security Blog
MongoDB | Blog
MongoDB | Blog
Cyberwarzone
Cyberwarzone
The Last Watchdog
The Last Watchdog
S
Securelist
N
News and Events Feed by Topic
S
Secure Thoughts
F
Fortinet All Blogs
博客园_首页
C
Cybersecurity and Infrastructure Security Agency CISA
量子位
M
MIT News - Artificial intelligence
F
Full Disclosure
T
The Blog of Author Tim Ferriss
T
Tailwind CSS Blog
Recent Commits to openclaw:main
Recent Commits to openclaw:main
Microsoft Security Blog
Microsoft Security Blog
I
InfoQ
P
Privacy International News Feed
L
LangChain Blog
Know Your Adversary
Know Your Adversary
C
CERT Recently Published Vulnerability Notes

Hacker News: Show HN

PurrrrrFocus: Pomodoro Timer App - App Store Workflow Engine — Multi-Step Orchestration for Bun RapidPhoto: Pro Photo Editor App - App Store GitHub - DheerG/swarms: Achieve extraordinary results with claude code across a variety of tasks SPICE simulation → oscilloscope → verification with Claude Code — Lucas Gerads Show HN: VCoding – A 5 MB native Windows IDE with no dynamic dependencies Show HN: LLMs don't hallucinate because they're bad at math, it's the format GitHub - Agent-FM/agentfm-core: AgentFM is a peer-to-peer network that turns everyday computers into a decentralized AI supercomputer. AgentFM lets you run massive AI workloads directly across a global mesh of idle CPUs and GPUs. Show HN: Tracking Top US Science Olympiad Alumni over Last 25 Years GitHub - Potarix/agent-hub: One place to talk to all your agents Show HN: Runtime security for AI agents(injection,tool abuse, data exfiltration) GitHub - dubeyKartikay/lazyspotify: Terminal Spotify client for macOS and Linux GitHub - the-banana-tool/king-louie: Easy to use GUI Personal AI Assistant. Win/Linux/Mac. Show HN I made my vacation rental bookable by AI agents–no Airbnb, 0% commission GitHub - basteez/jsf-autoreload: maven plugin to enable hot reload on jsf projects uvm32/hosts/host-gdbstub at main · ringtailsoftware/uvm32 GitHub - labsai/EDDI: Config-driven engine that turns JSON into production-grade AI agents. Multi-agent orchestration, 12+ LLM providers, MCP/A2A protocols, RAG, persistent memory, and enterprise compliance (EU AI Act, GDPR, HIPAA). Built on Quarkus. GitHub - glitchnsec/fortyone-oss: AI Executive Assistant Platform Quickstart | Alien GitHub - muxshed/shed: One stream in, or many. Every destination, simultaneously. No cloud middleman, no per-channel fees, no limits. GitHub - ocrbase-hq/ocrbase: 📄 PDF/IMG ->.MD/JSON Document OCR API for PaddleOCR and GLMOCR. Self-hostable. GitHub - impactjo/home-memory: MCP server that lets your AI assistant remember everything about your home. GitHub - Sets88/dbcls: DbCls is a powerful terminal database client that supports various databases GitHub - neptun2000/heor-agent-mcp GitHub - SeanFDZ/macmind: Single-layer transformer in HyperTalk for the classic Macintosh RollQuation: Math Puzzles - Apps on Google Play GitHub - dropbox/witchcraft Show HN: Agent-cache – Multi-tier LLM/tool/session caching for Valkey and Redis GitHub - opentalon/opentalon: OpenTalon is an open-source platform built from the ground up in Go as a robust alternative to OpenClaw LinkedIn™ 职位抓取工具 - Chrome 应用商店 GitHub - EdoardoBambini/Agent-Armor-Iaga: AI agents are getting tool access — shell, file system, databases, APIs, secrets. But **nobody is governing what they actually do with it**. Frameworks like LangChain, CrewAI, AutoGen, and Claude Code give agents the power to execute. Agent Armor gives you the power to control, audit, and approve every single action before it happens. HN Vibes — Week 15, Apr 7–13 2026 GitHub - chojs23/ec: Easy terminal-native 3-way git mergetool vim-like workflow GitHub - SethPyle376/hiraeth: Local AWS emulator focused on fast integration testing, with SQS support, SQLite-backed state, and a debug-friendly web UI. GitHub - JakOb-dotcom/cloud-sandbox-security-analysis: Technical analysis and Proof of Concept (PoC) regarding environment variable exfiltration in containerized cloud sandboxes via side-channel data leaks. Springboards - Flint Alpha Show HN: A simpler coding agent harness GitHub - audiodude/sudomake-friends GitHub - 256thFission/mini-mythos: OSS clone of Anthropic’s Mythos harness to locate C/C++ memory vulnerabilities Show HN: OpenParallax: OS-level privilege separation for AI agent execution Hacker News Sorted - Chrome 应用商店 Show HN: How to Install Docker on Ubuntu 24.04 LTS: Complete 2026 Guide GitHub - himanshudongre/smriti GitHub - sverrirsig/claude-control: macOS desktop dashboard for monitoring and managing multiple Claude Code sessions GitHub - ory/dockertest: Write better integration tests! Dockertest helps you boot up ephermal docker images for your Go tests with minimal work. Chiral - Chrome 应用商店 Show HN: Two Claudes collaborating through shared memory on a $100 mini-PC GitHub - pmichaillat/latex-cv: Minimalist LaTeX template for academic CVs GitHub - oguzbilgic/posse: A web UI for Anthropic Managed Agents. GitHub - sshiraz/depsly: Dependency risk analysis tool for npm packages ABI Add safari/agent-harness — Safari browser automation via safari-mcp by achiya-automation · Pull Request #212 · HKUDS/CLI-Anything GitHub - Halfblood-Prince/trustcheck: Verify PyPI package attestations and improve Python supply-chain security GitHub - oguzbilgic/kern-ai: Agents that do the work and show it. GitHub - bruits/satteri: High-performance Markdown and MDX processing for the JavaScript ecosystem GitHub - tylergibbs1/feedstock: High-performance web crawler and scraper for TypeScript, powered by Bun and Playwright GitHub - Grimm67123/grimmbot: The self-improving sandboxed and open-source AI agent. With persistent memory and scheduling. GitHub - whitevanillaskies/whitebloom: Local whiteboard that blooms. GitHub - hwdsl2/docker-whisper: Docker image for a self-hosted Whisper speech-to-text server with speaker diarization and OpenAI-compatible transcription and translation APIs. Powered by faster-whisper. Supports all Whisper models, NVIDIA GPU (CUDA) acceleration, JSON/SRT/VTT output, SSE streaming, offline mode, and multi-arch (amd64, arm64). GitHub - yisding/reviewwiggum GitHub - MarwanAlsoltany/serrors: Structured errors for Go: sentinel hierarchies, typed data, custom formatting, and slog integration. GitHub - soatok/age-php GitHub - Luthiraa/markitme GitHub - stagas/rtdiff: realtime git diff gui and AI-assisted commits GitHub - tombedor/excalicharts GitHub - wh1le/excalidraw-edit: Open and edit .excalidraw files from the terminal. Offline, auto-saves to disk. MalExt Sentry - Malicious Extension Scanner - Chrome 应用商店 GitHub - syi0808/asciianimesvg: Generate animated ASCII art SVGs from text. CLI, Rust library, WASM, and web editor. GitHub - zaina-ml/ml_forge: A visual-based graph node editor for training computer vision models. GitHub - anakin87/llm-rl-environments-lil-course: 🌱 A little course on Reinforcement Learning Environments for evaluating and training Language Models GitHub - takaakit/superpowers-uml: Superpowers-UML modifies Superpowers to ensure a software development workflow in which AI agents design through UML modeling. AdriByte Studio - Sviluppo Web e Soluzioni Digitali GitHub - chouligi/angel-copilot: Your personalized Angel Investment Advisor Show HN: MoodSense AI (ML and FastAPI and Gradio, Deployed on Hugging Face) Moodsense Ai - a Hugging Face Space by aman179102 GitHub - agenteractai/lodmem: Level Of Detail Context Management for Agents GitHub - ostefani/subnetlens: A fast, concurrent network scanner with a TUI and plain-text CLI, built in Go. It discovers live hosts on your network, scans their open ports, resolves hostnames, and fingerprints operating systems—delivered. Cyber Pulse: Agentic Intel - Apps on Google Play Whisper API: Self-Hostable Speech to Text Transcription The Agent-Web Protocol Stack: A Research Thesis GitHub - msmarkgu/RelayFreeLLM: A restful API designed to route user prompts to various AI model providers. Show HN: Provepy – A Python decorator that proves your code using Lean and LLMs Show HN: Pardonned.com – A searchable database of US Pardons GitHub - patrickdappollonio/dux: Dux is a terminal UI that lets you run multiple AI coding agents side by side, each in its own git worktree, with full companion terminals, macros, commit generation, and a command palette that knows more tricks than you do. kMC Crystal Simulator Show HN: HyperFlow – A self-improving agent framework built on LangGraph GitHub - stef41/vibescore: 🎵 Grade your vibe-coded project. One command, instant letter grade across security, quality, dependencies, and testing. GitHub - stef41/lmscan: 🔍 Detect AI-generated text and fingerprint which LLM wrote it. Open-source GPTZero alternative. Zero dependencies, works offline. imgur.com GitHub - visionscaper/collabmem: Enabling long-term collaboration with Agentic AI - building up episodic and world model memory over time with in-context awareness 在 Steam 上购买 FriedrichAI: Offline AI 立省 10% GitHub - atripati/ark: AI Runtime Kernel — a context operating system for AI agents. Eliminates tool bloat, loads only what’s needed, and gives LLMs their reasoning space back. GitHub - nowork-studio/toprank: Open-source Claude Code skills for SEO, SEM, Google Ads GitHub - tacomanator/sash: Lightweight macOS menu bar app for reliably cycling through windows of the current application. Appents | Social Media Management for Product-First Teams GitHub - pnhoang/youtube-spam-blocker: Automatically detects and hides spam messages in YouTube Live chat. Set rate limits, keyword filters, and block repeat offenders. GitHub - decisionnode/DecisionNode: CLI + Local MCP - A shared structured memory store across Claude Code, Cursor, Windsurf, Antigravity, and every MCP client. Semantically queryable. GitHub - AvaCodeSolutions/django-email-learning: An open source Django app for creating email-based learning platforms with IMAP integration and React frontend components. The $100K Gap in Kubernetes Security Tooling Function Calling Harness: From 6.75% to 100%
GitHub - JinBa1/my-time-has-come: A local CLI safety brake that instruct Claude Code to wrap up before the 5-hour usage limit runs out.
BuyG1n · 2026-05-15 · via Hacker News: Show HN

My Time Has Come banner

An automatic safety brake for Claude Code before your usage windows run out

CI Go Report Card Go 1.26.2 License: MIT Release: v0.3.0

Ever found yourself watching the usage bar instead of the work? Or burned through a usage window mid-task and returned to a messy, half-edited repo?

mthc gives Claude Code a two-stage stop mechanism. You set soft and hard thresholds for "wrap up as soon as you can" and "stop using tools".

At the soft threshold, mthc injects a customizable wrap-up prompt so Claude can summarize progress, stabilize the repo, and write a handoff in the style you prefer. At the hard threshold, mthc blocks further tool use through Claude Code hooks and writes a deterministic handoff report, so the next session has a clear restart point.

mthc is the emergency brake you barely notice until it matters: Install it once, then keep using Claude Code normally. It stays quiet in the background and only steps in when the time comes.

  • Drop in and keep working: no alternate workflow, no new session manager.
  • Claude Code native: uses Claude Code statusline and hook surfaces directly; no daemon or runtime dependency.
  • Soft landing before cutoff: customizable wrap-up prompts and handoff files.
  • Hard stop when needed: tool-use blocking plus a deterministic restart report.
  • Local and lightweight: local state only, no daemon, no telemetry, no network calls.
  • Multi-session aware: one local source of truth across concurrent sessions.

Contents

  • Quick Start
    • Install
    • Configure Thresholds
    • Customize Handoffs
  • How It Works
    • How Claude Responds To Soft Stop
    • How Claude Responds To Hard Stop
    • Architecture
    • Program Boundaries
  • Commands
  • Troubleshooting
  • Roadmap
  • Contributing
  • License

Quick Start

Install

Requirements:

  • Claude Code CLI
  • Linux-style or macOS shell environment
  • Node.js and npm for the preferred npm install path
  • mthc available on the PATH seen by Claude Code

The preferred install path is npm.

npm install example
npm install -g @jinba1/mthc
mthc install
mthc doctor
mthc status
GitHub Releases archive example
curl -LO https://github.com/JinBa1/my-time-has-come/releases/download/v0.3.0/mthc_0.3.0_linux_amd64.tar.gz
tar -xzf mthc_0.3.0_linux_amd64.tar.gz
sudo install -m 0755 mthc /usr/local/bin/mthc
mthc install
mthc doctor

Use the archive that matches your operating system and CPU. Release packages are built for Linux and macOS targets.

Source build fallback requires Go 1.26.2.

Source build fallback
git clone https://github.com/JinBa1/my-time-has-come.git
cd my-time-has-come
go build -o mthc .
sudo install -m 0755 mthc /usr/local/bin/mthc
mthc install
mthc doctor
mthc status

mthc install updates ~/.claude/settings.json, writes ~/.config/mthc/config.toml, and initializes ~/.config/mthc/state.json. It captures your prior Claude Code statusline command so uninstall can restore it later.

Uninstall:

  • removes mthc hook/statusline entries
  • restores the captured prior statusline where possible

Configure Thresholds

Show the resolved config:

Defaults:

[policy]
enabled = true # Set false to observe status but disable soft/hard interventions.

[thresholds.five_hour]  # customize your own thresholds.
enabled = true
soft_pct = 85 # Ask Claude to wrap up when 5-hour usage reaches this percentage.
hard_pct = 95 # Arm the hard-stop gate and deterministic handoff at this percentage.

[thresholds.seven_day]
enabled = true
soft_pct = 90 # Weekly quota is larger.
hard_pct = 98

[handoff]
path_template = "{cwd}/.mthc/handoff-{session_id}-{window_id}-{window_start_ts}.md" # Where handoff files are written.

[display]
mode = "silent" # Keep mthc quiet and preserve your existing Claude Code statusline.

[statusline]
refresh_interval_seconds = 10 # Session liveness interval; active sessions are tracked at about 2x this value.

[hard_stop]
enable_pretool_deny = true # Block future tool calls after the hard threshold is reached.

Change thresholds with:

mthc config set thresholds.five_hour.soft_pct 80
mthc config set thresholds.seven_day.enabled false
mthc config set policy.enabled false
mthc config validate

Config resolution is:

built-in defaults < ~/.config/mthc/config.toml < ./.mthc/config.toml

Do not hand-edit the [internal] section in ~/.config/mthc/config.toml. It is install metadata used to restore your prior Claude Code settings.

Customize Handoffs

At the soft threshold, mthc injects a wrap-up prompt through Claude Code's PostToolBatch hook. The default prompt asks the agent to summarize the task, record repository state, write a handoff document, and stop.

You can provide your own soft prompt template:

mthc config set handoff.soft_prompt_path /path/to/soft-prompt.tmpl

The soft prompt template uses Go text/template variables:

{{.WindowID}} {{.WindowLabel}}
{{.UsedPercentage}} {{.ResetsAtHuman}} {{.ResetsAtUnix}}
{{.FiveHour.UsedPercentage}} {{.FiveHour.ResetsAtHuman}} {{.FiveHour.ResetsAtUnix}} {{.FiveHour.Absent}}
{{.SevenDay.UsedPercentage}} {{.SevenDay.ResetsAtHuman}} {{.SevenDay.ResetsAtUnix}} {{.SevenDay.Absent}}
{{.SessionID}} {{.HandoffPath}} {{.CWD}} {{.ModelID}}

For custom soft prompts, .UsedPercentage, .ResetsAtHuman, and .ResetsAtUnix describe the trigger window. If an existing template hard-codes 5-hour wording, use .WindowLabel in the copy, or use .FiveHour.* / .SevenDay.* when the prompt needs specific windows.

At the hard threshold, mthc writes a deterministic Markdown handoff before denying future tool use. The handoff includes:

  • session ID and model ID
  • trigger window, usage percentage, and reset time
  • working directory
  • transcript path
  • resume instructions for the next Claude Code session

The default hard-stop handoff path is:

{cwd}/.mthc/handoff-{session_id}-{window_id}-{window_start_ts}.md

How It Works

How Claude Responds To Soft Stop

The soft stop appears inside the normal Claude Code conversation. In this example, Claude finishes reading context, acknowledges the usage-window warning, and writes the requested handoff file instead of continuing with the original plan.

Claude Code responding to an mthc soft stop

How Claude Responds To Hard Stop

The hard stop is more forceful. In this example, Claude tries to write another file after the gate is armed, and Claude Code reports the PreToolUse blocking error instead of allowing the edit.

Claude can still respond in text, so the final message points to the reset time and the handoff path rather than pretending the process was killed.

Claude Code responding to an mthc hard stop

Architecture

                Claude Code
             ┌──────┴────────┐
             │               │
        statusLine       hooks
         command    PostToolBatch / PreToolUse
             │               │
             ▼               ▼
   ┌────────────────┐  ┌────────────────┐
   │ statusline-shim│  │   hook-shim    │
   └───────┬────────┘  └───────┬────────┘
           │                   │
           │ observes usage    │ applies policy
           │ preserves prior   │ soft prompt / deny
           │ statusline        │
           └─────────┬─────────┘
                     ▼
        ┌──────────────────────────┐
        │ ~/.config/mthc/state.json│
        │ config.toml              │
        └────────────┬─────────────┘
                     │
         ┌───────────┴───────────┐
         ▼                       ▼
 soft wrap-up prompt    hard gate + handoff

The statusline shim observes Claude Code's usage-window payload and updates local state. The hook shim applies the policy when Claude Code invokes PostToolBatch or PreToolUse.

This split matters: the hard gate is armed during statusline processing, then enforced on later tool-use hooks. That keeps a crossed threshold from being lost just because a later observation is transiently lower.

Program Boundaries

  • Soft stop is advisory. The model decides whether and when to follow it.
  • Hard stop blocks future tool calls. It does not kill Claude Code, interrupt an in-flight tool, erase context already loaded by the model, or prevent a final text response.
  • Both controls depend on Claude Code hook opportunities.
  • State is local to one machine.

Commands

Command Purpose
mthc install Install the statusline and hook shims into Claude Code settings
mthc uninstall Remove mthc hooks and restore prior settings
mthc status Show current policy, usage-window, session, and handoff state
mthc doctor Diagnose installation and runtime problems
mthc config show Print resolved configuration
mthc config set thresholds.five_hour.soft_pct 80 Update a user config value
mthc config set thresholds.seven_day.enabled false Disable one policy window
mthc config set policy.enabled false Disable policy decisions
mthc config validate Validate the user config file
mthc dismiss --soft Clear soft-injection state for the active window
mthc dismiss --hard Disarm the hard gate until the next observation re-arms it
mthc record start Capture statusline and hook events as JSONL
mthc playback replay <dir> Replay captured events through the core policy path

Troubleshooting

mthc is designed to fail open. If a hook cannot parse input, read config, lock state, or render output, it emits an empty response so Claude Code is not blocked by a broken governor.

Symptom First check
mthc is not found by Claude Code Confirm the installed binary is on the PATH Claude Code sees
Hooks do not appear to run Run mthc doctor and inspect the reported Claude Code settings state
Prior statusline disappeared Run mthc uninstall; it restores the captured command when config metadata is intact
Hard gate did not interrupt a running tool Expected; the gate applies on later PreToolUse calls
Need a reproducible report Run mthc record start, reproduce, then mthc record stop

When something behaves strangely, you can record the hook/statusline events and replay the policy decisions without touching live Claude Code state.

Recording files can contain local paths and session metadata. Treat recordings as private unless you have inspected them.

Roadmap

  • Initial release supporting Claude Code CLI
  • Packaged binaries and npm distribution
  • Codex CLI, GEMINI CLI and OpenCode adapters
  • Smarter thresholding based on burn rate and time-to-reset
  • Stronger interruption mechanisms as Claude Code control surfaces evolve

Contributing

Issues and focused pull requests are welcome.

Before opening a pull request, run:

env GOCACHE=/tmp/mthc-go-build go build ./...
env GOCACHE=/tmp/mthc-go-build go vet ./...
env GOCACHE=/tmp/mthc-go-build go test ./...
gofmt -l .

License

This project is licensed under the MIT license.