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

WordPress大学
WordPress大学
Cyberwarzone
Cyberwarzone
The GitHub Blog
The GitHub Blog
云风的 BLOG
云风的 BLOG
P
Proofpoint News Feed
小众软件
小众软件
Recent Announcements
Recent Announcements
博客园 - 三生石上(FineUI控件)
Security Archives - TechRepublic
Security Archives - TechRepublic
W
WeLiveSecurity
Cloudbric
Cloudbric
博客园 - 司徒正美
美团技术团队
N
News and Events Feed by Topic
Exploit-DB.com RSS Feed
Exploit-DB.com RSS Feed
PCI Perspectives
PCI Perspectives
宝玉的分享
宝玉的分享
H
Help Net Security
H
Hackread – Cybersecurity News, Data Breaches, AI and More
Google DeepMind News
Google DeepMind News
Help Net Security
Help Net Security
Last Week in AI
Last Week in AI
S
Schneier on Security
N
News | PayPal Newsroom
B
Blog RSS Feed
L
LINUX DO - 最新话题
T
Troy Hunt's Blog
S
Secure Thoughts
雷峰网
雷峰网
aimingoo的专栏
aimingoo的专栏
L
Lohrmann on Cybersecurity
G
Google Developers Blog
Microsoft Azure Blog
Microsoft Azure Blog
Hacker News - Newest:
Hacker News - Newest: "LLM"
T
Tenable Blog
S
Securelist
L
LangChain Blog
Recent Commits to openclaw:main
Recent Commits to openclaw:main
I
InfoQ
H
Heimdal Security Blog
Cisco Talos Blog
Cisco Talos Blog
F
Full Disclosure
Y
Y Combinator Blog
酷 壳 – CoolShell
酷 壳 – CoolShell
K
Kaspersky official blog
T
Tailwind CSS Blog
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
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
阮一峰的网络日志
阮一峰的网络日志
C
Cisco Blogs

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 - apvcode/PromptShark: Stop burning tokens on infinite loops. An interactive proxy and debugger for AI agents with Time-Travel capabilities (built with Go & C++).
apvcode · 2026-06-17 · via Hacker News: Show HN

MITM proxy debugger & loop detector for AI Agents

Go C++ SQLite License

Intercept every prompt. Catch every loop. Replay any step.

PromptShark Dashboard

🌍 English · Русский


🇬🇧 English

What is PromptShark?

PromptShark is a transparent local proxy that sits between your AI agent and the OpenAI API. It captures, logs, and visualizes every single request/response pair in a beautiful real-time dashboard — and automatically detects when your agent gets stuck in an infinite tool-calling loop.

One line to integrate. Zero code changes to your agent.
Just swap base_urlhttp://localhost:8080/v1


At a Glance

🔍 Intercept & Inspect

Every API call flows through PromptShark. View structured request/response pairs, collapse raw JSON payloads, and track token usage with per-step Prompt / Completion breakdowns.

🚨 Loop Detection (C++)

A dedicated C++ engine runs via IPC alongside the Go proxy. It hashes every tool call and flags repeating patterns instantly — before they burn through your API balance.

⏪ Time-Travel Replay

Click on any past step, edit the JSON payload in the built-in editor, preview your changes with an interactive diff view, and re-run the agent from that point. Earlier steps are served from the SQLite cache — no extra API costs.

📊 Live Dashboard

Dark-themed glassmorphism UI with WebSocket-powered live streaming. Auto-switching sessions, instant step search, Markdown export, and session management — all in a single embedded HTML page.


Feature Matrix

Category Feature Details
Proxy Drop-in integration Change one URL, keep your existing OpenAI SDK
Proxy Streaming + non-streaming Full support for stream: true and stream: false
Proxy Token tracking Per-step prompt_tokens / completion_tokens split
Proxy USD Cost Tracker Real-time per-step and per-session cost in dollars
Proxy Latency measurement Time-to-first-token (TTFT) displayed on each step
Detection C++ Loop Detector IPC-connected engine with configurable threshold
Detection Configurable sensitivity loop_config.txt — threshold & argument-ignoring
Debug Time-Travel Replay Edit & re-run from any step using cached responses
Debug Interactive Diff View Side-by-side diff of original vs. edited payload
Debug Raw JSON Viewer Collapsible request/response payloads per step
Testing Chaos Mode Toggle to inject fake 429 Rate Limit errors from UI
UI Real-time WebSockets Tokens stream live into the dashboard
UI Session search & filter Instant full-text search across all steps
UI Auto-switching sessions Dashboard follows the active agent run
Export Markdown reports One-click export of entire debug sessions
Export Session management Create, switch, and delete sessions from sidebar

Installation

There are three ways to run PromptShark:

Option A: Docker (Recommended)

Option B: Pre-compiled Binaries

Download the latest .zip or .tar.gz for your OS (Windows, macOS, Linux) from the GitHub Releases page. Extract and run agent_supervisor (or agent_supervisor.exe).

Option C: Build from Source

1 · Build the C++ Core

mkdir -p build && cd build
cmake .. && make

2 · Run the Proxy

cd proxy
go build -o ../build/agent_supervisor .
../build/agent_supervisor

You will see:

    ____                       __  _____ __               __  
   / __ \_________  ____ ___  / /_/ ___// /_  ____ ______/ /__
  / /_/ / ___/ __ \/ __ '__ \/ __ \__ \/ __ \/ __ '/ ___/ //_/
 / ____/ /  / /_/ / / / / / / /_/ /__/ / / / / /_/ / /  / ,<   
/_/   /_/   \____/_/ /_/ /_/ .___/____/_/ /_/\__,_/_/  /_/|_|  
                          /_/                                   
  ──────────────────────────────────────────────────
  🦈 Version     : 0.1.0
  🌐 Dashboard   : http://localhost:8080
  📦 Database    : ../db/promptshark.db
  🎯 Target API  : https://api.openai.com
  🔗 Proxy URL   : http://localhost:8080/v1
  ──────────────────────────────────────────────────

  Point your OpenAI SDK base_url to the Proxy URL above.
  Press Ctrl+C to stop.

Open http://localhost:8080 — the dashboard is ready.

3 · Point your Agent

Python
from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:8080/v1",
    api_key="sk-..."  # forwarded transparently to OpenAI
)

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}],
    stream=True
)
Node.js
const OpenAI = require("openai");
const client = new OpenAI({
  baseURL: "http://localhost:8080/v1",
  apiKey: "sk-..."
});

const stream = await client.chat.completions.create({
  model: "gpt-4o",
  messages: [{ role: "user", content: "Hello!" }],
  stream: true,
});

Loop Detector Config

Edit loop_config.txt in the project root:

# Consecutive identical calls before triggering alert
THRESHOLD=3

# Ignore arguments — match function name only
IGNORE_ARGS=false
Parameter Default Effect
THRESHOLD 3 Number of identical consecutive tool calls that trigger a loop alarm
IGNORE_ARGS false When true, only function names are compared (arguments are stripped)

⚙️ CLI Options

Flag Default Description
--port 8080 HTTP server port
--db ../db/promptshark.db Path to SQLite database file
--schema ../db/schema.sql Path to schema.sql file
--target https://api.openai.com Upstream LLM API URL

Environment variables PORT and DB_PATH are also supported (flags take priority).

The server handles graceful shutdown: press Ctrl+C and it will finish in-flight requests, close the database, and exit cleanly.


Tech Stack

Layer Technology Role
Proxy Server Go 1.22+ HTTP reverse proxy, WebSocket hub, REST API, CLI
Loop Engine C++ 17 Real-time tool-call hashing & pattern detection via stdin/stdout IPC
Storage SQLite (WAL) Embedded session & step persistence, zero-config
Dashboard HTML + Tailwind + JS Single-file embedded UI served by Go via go:embed
Diff Engine jsdiff Client-side line-by-line diff rendering for Time-Travel

Roadmap

  • Configurable loop rules via YAML (regex patterns, token budgets)
  • Ollama / Anthropic Claude API support
  • Multi-agent message flow visualizer (network graph)

🇷🇺 Русский

Что такое PromptShark?

PromptShark — это прозрачный локальный прокси, который встает между вашим ИИ-агентом и API OpenAI. Он перехватывает, логирует и визуализирует каждую пару запрос/ответ в красивом дашборде в реальном времени — и автоматически обнаруживает, если агент попал в бесконечный цикл вызова инструментов.

Одна строка для интеграции. Ноль изменений в коде агента.
Просто замените base_urlhttp://localhost:8080/v1


Обзор возможностей

🔍 Перехват и инспекция

Каждый API-вызов проходит через PromptShark. Смотрите структурированные пары запрос/ответ, раскрывайте сырой JSON и отслеживайте использование токенов с разбивкой Prompt / Completion для каждого шага.

🚨 Детектор циклов (C++)

Выделенное C++ ядро работает через IPC параллельно с Go прокси. Оно хэширует каждый вызов инструмента и мгновенно фиксирует повторяющиеся паттерны — до того, как они сожгут ваш бюджет API.

⏪ Машина времени (Replay)

Кликните на любой предыдущий шаг, отредактируйте JSON в встроенном редакторе, проверьте изменения в интерактивном diff-просмотрщике и перезапустите агента с этой точки. Предыдущие шаги отдаются из кэша SQLite — никаких лишних затрат.

📊 Живой дашборд

Темная тема с эффектом стекла, стриминг по WebSocket, авто-переключение между сессиями, мгновенный поиск по шагам, экспорт в Markdown и управление сессиями — всё в одном встроенном HTML-файле.


Матрица возможностей

Категория Функция Подробности
Прокси Drop-in интеграция Измените один URL, оставьте текущий OpenAI SDK
Прокси Стриминг + обычные запросы Полная поддержка stream: true и stream: false
Прокси Трекинг токенов Разбивка prompt_tokens / completion_tokens на каждый шаг
Прокси Калькулятор стоимости Цена каждого шага и всей сессии в долларах в реальном времени
Прокси Замер задержки Время ответа API (TTFT) на каждом шаге
Детекция C++ Loop Detector Движок через IPC с настраиваемым порогом
Детекция Гибкая настройка loop_config.txt — порог срабатывания и игнорирование аргументов
Отладка Time-Travel Replay Редактируй и перезапускай с любого шага через кэш
Отладка Интерактивный Diff Визуальное сравнение оригинала и отредактированного промпта
Отладка Raw JSON Viewer Сворачиваемые JSON запроса/ответа для каждого шага
Тестирование Chaos Mode Тумблер для инъекции фейковых 429 Rate Limit ошибок из UI
UI WebSocket реалтайм Токены стримятся в дашборд в реальном времени
UI Поиск и фильтрация Мгновенный полнотекстовый поиск по всем шагам
UI Авто-переключение Дашборд следует за активным запуском агента
Экспорт Markdown отчёты Экспорт всей сессии одним кликом
Экспорт Управление сессиями Создание, переключение и удаление из сайдбара

Установка

Есть три способа запустить PromptShark:

Вариант А: Docker (Рекомендуется)

Вариант Б: Готовые бинарники

Скачайте актуальный .zip или .tar.gz для вашей ОС (Windows, macOS, Linux) со страницы GitHub Releases. Распакуйте и запустите agent_supervisor (или agent_supervisor.exe).

Вариант В: Сборка из исходников

1 · Сборка C++ ядра

mkdir -p build && cd build
cmake .. && make

2 · Запуск прокси

cd proxy
go build -o ../build/agent_supervisor .
../build/agent_supervisor

Вы увидите:

    ____                       __  _____ __               __  
   / __ \_________  ____ ___  / /_/ ___// /_  ____ ______/ /__
  / /_/ / ___/ __ \/ __ '__ \/ __ \__ \/ __ \/ __ '/ ___/ //_/
 / ____/ /  / /_/ / / / / / / /_/ /__/ / / / / /_/ / /  / ,<   
/_/   /_/   \____/_/ /_/ /_/ .___/____/_/ /_/\__,_/_/  /_/|_|  
                          /_/                                   
  ──────────────────────────────────────────────────
  🦈 Version     : 0.1.0
  🌐 Dashboard   : http://localhost:8080
  📦 Database    : ../db/promptshark.db
  🎯 Target API  : https://api.openai.com
  🔗 Proxy URL   : http://localhost:8080/v1
  ──────────────────────────────────────────────────

  Point your OpenAI SDK base_url to the Proxy URL above.
  Press Ctrl+C to stop.

Откройте http://localhost:8080 — дашборд готов к работе.

3 · Подключите агента

Python
from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:8080/v1",
    api_key="sk-..."  # прозрачно перенаправляется в OpenAI
)

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Привет!"}],
    stream=True
)
Node.js
const OpenAI = require("openai");
const client = new OpenAI({
  baseURL: "http://localhost:8080/v1",
  apiKey: "sk-..."
});

const stream = await client.chat.completions.create({
  model: "gpt-4o",
  messages: [{ role: "user", content: "Привет!" }],
  stream: true,
});

Настройка детектора циклов

Отредактируйте loop_config.txt в корне проекта:

# Количество идентичных вызовов подряд до срабатывания тревоги
THRESHOLD=3

# Игнорировать аргументы — сравнивать только имя функции
IGNORE_ARGS=false
Параметр По умолчанию Эффект
THRESHOLD 3 Сколько одинаковых последовательных вызовов считать зацикливанием
IGNORE_ARGS false Если true, сравниваются только имена функций (аргументы игнорируются)

⚙️ Параметры CLI

Флаг По умолчанию Описание
--port 8080 Порт HTTP-сервера
--db ../db/promptshark.db Путь к файлу базы данных SQLite
--schema ../db/schema.sql Путь к файлу схемы
--target https://api.openai.com URL upstream LLM API

Также поддерживаются переменные окружения PORT и DB_PATH (флаги имеют приоритет).

Сервер поддерживает Graceful Shutdown: нажмите Ctrl+C, и он корректно завершит текущие запросы, закроет базу данных и выведет ✅ Goodbye!.


Технологический стек

Слой Технология Роль
Proxy Server Go 1.22+ HTTP reverse proxy, WebSocket хаб, REST API, CLI
Loop Engine C++ 17 Хэширование и детекция паттернов через stdin/stdout IPC
Хранилище SQLite (WAL) Встроенная база сессий и шагов, без настройки
Дашборд HTML + Tailwind + JS Единый встроенный UI через go:embed
Diff Engine jsdiff Клиентский построчный diff для Time-Travel

Дорожная карта

  • Конфигурация правил детектора через YAML (regex паттерны, бюджеты токенов)
  • Поддержка Ollama / Anthropic Claude API
  • Визуализатор потоков мультиагентных систем (граф сети)

[MIT License](LICENSE) — use it, fork it, ship it.