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

T
Threatpost
The Hacker News
The Hacker News
AWS News Blog
AWS News Blog
Spread Privacy
Spread Privacy
T
Tenable Blog
C
CERT Recently Published Vulnerability Notes
Cisco Talos Blog
Cisco Talos Blog
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
S
Securelist
P
Privacy & Cybersecurity Law Blog
Know Your Adversary
Know Your Adversary
T
The Exploit Database - CXSecurity.com
Latest news
Latest news
D
Darknet – Hacking Tools, Hacker News & Cyber Security
I
Intezer
F
Fortinet All Blogs
Engineering at Meta
Engineering at Meta
Simon Willison's Weblog
Simon Willison's Weblog
The Register - Security
The Register - Security
CTFtime.org: upcoming CTF events
CTFtime.org: upcoming CTF events
L
Lohrmann on Cybersecurity
C
Cyber Attacks, Cyber Crime and Cyber Security
Microsoft Azure Blog
Microsoft Azure Blog
P
Proofpoint News Feed
H
Help Net Security
T
Threat Research - Cisco Blogs
D
DataBreaches.Net
S
Schneier on Security
Cyberwarzone
Cyberwarzone
Google DeepMind News
Google DeepMind News
P
Privacy International News Feed
S
Secure Thoughts
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
Recorded Future
Recorded Future
C
Cybersecurity and Infrastructure Security Agency CISA
MyScale Blog
MyScale Blog
M
MIT News - Artificial intelligence
Stack Overflow Blog
Stack Overflow Blog
IT之家
IT之家
人人都是产品经理
人人都是产品经理
NISL@THU
NISL@THU
博客园 - Franky
T
Tor Project blog
G
GRAHAM CLULEY
博客园 - 【当耐特】
Jina AI
Jina AI
Security Archives - TechRepublic
Security Archives - TechRepublic
K
KPMG report finds enterprise disconnect between AI and its ROI | CIO
A
About on SuperTechFans
Hacker News - Newest:
Hacker News - Newest: "LLM"

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 - m-ghalib/gemini-plugin-cc: Gemini plugin for Claude Code - reviews, and adversarial reviews to Google's Gemini CLI
morawr · 2026-04-22 · via Hacker News: Show HN

Use Gemini from inside Claude Code for code reviews or to delegate tasks to Gemini.

This plugin lets you delegate tasks, reviews, and adversarial reviews to Google's Gemini CLI (@google/gemini-cli@0.38.2) without leaving your Claude Code session. It wraps the Gemini CLI via the Agent-Client Protocol (ACP) -- JSON-RPC 2.0 over stdio -- so Claude and Gemini can work side by side on the same codebase.


What You Get

Six slash commands and a proactive subagent:

Command Purpose
/gemini:review Code review with structured findings
/gemini:adversarial-review Steerable adversarial review that challenges design choices
/gemini:rescue Delegate investigation, fix, diagnosis, research, or any substantial task to Gemini
/gemini:status Check active and recent Gemini jobs
/gemini:result View the stored output of a finished job
/gemini:cancel Cancel an active background job
/gemini:setup Verify readiness, check auth, toggle review gate

gemini-rescue subagent -- Claude Code can proactively delegate substantial tasks to Gemini through this subagent. It handles debugging, implementation, research, writing, analysis, brainstorming, explanation, or a second diagnosis pass as a thin forwarding wrapper around the Gemini companion runtime.


Requirements

  • Google account with Gemini CLI access, or a Gemini API key, or Vertex AI credentials
  • Node.js 18.18 or later
  • Gemini CLI (@google/gemini-cli@0.38.2) -- the setup command can install it for you

Install

/plugin marketplace add m-ghalib/gemini-plugin-cc
/plugin install gemini@gemini-plugin-cc
/reload-plugins
/gemini:setup

The /gemini:setup command checks whether the Gemini CLI is installed and authenticated. If the CLI is missing and npm is available, it offers to install it for you via npm install -g @google/gemini-cli@0.38.2.


Usage

/gemini:review

Reads your working-tree diff or branch diff and asks Gemini to review it. Returns structured findings covering correctness, regression risk, and code quality. Review-only -- never modifies code.

Flags: --base <ref> (explicit base branch), --scope auto|working-tree|branch, --wait (foreground), --background (detach).

If neither --wait nor --background is specified, Claude estimates the review size and recommends an execution mode.

/gemini:review
/gemini:review --base main
/gemini:review --scope branch --background

/gemini:adversarial-review

Same target selection as /gemini:review, but challenges the chosen implementation, design choices, tradeoffs, and assumptions. Not just a stricter pass over defects -- it asks whether the current approach is the right one and where the design could fail under real-world conditions.

Steerable: accepts focus text after flags to direct the review toward specific concerns.

Flags: same as /gemini:review.

/gemini:adversarial-review
/gemini:adversarial-review --base main challenge the error handling strategy
/gemini:adversarial-review challenge whether this retry logic handles all failure modes

/gemini:rescue

Delegates code investigation, fixes, or diagnosis to Gemini. For exploratory or diagnostic tasks, the plugin suggests that Gemini use its built-in codebase_investigator subagent before drafting a fix.

Flags:

Flag Description
--background Detach and run as a background job
--wait Run in the foreground (block until complete)
--resume Continue from the last Gemini thread in this repository
--fresh Start a new Gemini thread (ignore any prior session)
--model <alias> Choose a specific Gemini model (see model aliases below)
--effort low|medium|high Set Gemini's thinking level (accepted; pending upstream ACP support)
--plan Read-only planning run (Gemini proposes but does not execute)

Model aliases:

Alias Resolves to
pro gemini-3.1-pro-preview
pro-3 gemini-3.1-pro-preview
flash gemini-3-flash-preview
flash-lite gemini-3.1-flash-lite-preview
2.5-pro gemini-2.5-pro
2.5-flash gemini-2.5-flash
2.5-flash-lite gemini-2.5-flash-lite
auto auto-gemini-3
auto-2.5 auto-gemini-2.5

Concrete model IDs are also accepted (e.g., --model gemini-3-pro-preview).

Examples:

/gemini:rescue investigate why the auth middleware returns 403 for valid tokens
/gemini:rescue --background fix the flaky test in user-service
/gemini:rescue --plan trace the data flow from API to database for the orders endpoint
/gemini:rescue --model flash --effort high diagnose the memory leak in the worker pool

/gemini:status

Shows active and recent Gemini jobs for the current repository.

Flags:

Flag Description
--wait Poll until a job finishes; auto-resolves if one active job exists, otherwise pass a job ID
--timeout-ms <n> Maximum time to poll (used with --wait)
--all Include jobs from all sessions, not just the current one

Examples:

/gemini:status
/gemini:status --wait
/gemini:status --all

/gemini:result

Shows the stored output of a finished Gemini job. Preserves all details including the session ID (useful for --resume on follow-up commands).

Accepts an optional job ID. Defaults to the latest finished job.

Examples:

/gemini:result
/gemini:result abc123

/gemini:cancel

Cancels an active background Gemini job by sending an ACP cancel request to the running session.

Accepts an optional job ID. Defaults to the latest active job.

Examples:

/gemini:cancel
/gemini:cancel abc123

/gemini:setup

Checks Gemini CLI availability and authentication status. Reports the detected auth method (OAuth, API key, Vertex AI, gateway) and whether credentials are present locally. Use --verify to confirm credentials work end-to-end.

If the Gemini CLI is not installed and npm is available, offers to install it via npm install -g @google/gemini-cli.

Flags:

Flag Description
--verify Confirm credentials work end-to-end via ACP handshake (requires Gemini CLI)
--enable-review-gate Enable the stop-time review gate (Gemini reviews Claude's work before stopping)
--disable-review-gate Disable the stop-time review gate

The review gate is a Stop hook that blocks Claude from finishing until Gemini has reviewed the work. When the gate cannot reach Gemini (auth failure, network issue), it blocks and surfaces actionable guidance rather than silently allowing the stop. Disable with --disable-review-gate if the block is not resolvable in the current session.

Limitation: The gate receives Claude's last response text and the current repository state, not a per-turn diff. In multi-turn sessions with a dirty working tree, Gemini uses the response text to attribute changes to the most recent turn, but cannot perfectly distinguish this turn's edits from earlier ones.

Examples:

/gemini:setup
/gemini:setup --enable-review-gate
/gemini:setup --disable-review-gate

Typical Flows

Quick code review:

/gemini:review

Deep adversarial review of a feature branch:

/gemini:adversarial-review --base main challenge whether this retry logic handles all failure modes

Delegate a complex investigation:

/gemini:rescue --background investigate the N+1 query problem in the user dashboard
/gemini:status --wait
/gemini:result

Continue a prior Gemini thread:

/gemini:rescue --resume apply the top fix from your investigation

General-purpose delegation via rescue:

/gemini:rescue research the best rate limiting libraries for Node.js and recommend one

Gemini Integration

The plugin communicates with the Gemini CLI through the Agent-Client Protocol (ACP) -- a JSON-RPC 2.0-based protocol over stdio. The Gemini CLI is started with gemini --acp, which launches it in programmatic ACP mode rather than the interactive terminal UI. ACP supports session management, turn-based prompting, streaming updates, and session mode control.

Authentication

Configure your auth method in ~/.gemini/settings.json. Supported auth types:

Auth type Configuration
Google OAuth Run gemini interactively to complete OAuth flow, or set selectedType: "oauth-personal"
Gemini API key Set selectedType: "gemini-api-key" and GEMINI_API_KEY env var
Google API key Set selectedType: "google-api-key" and GOOGLE_API_KEY env var
Vertex AI Set selectedType: "vertex-ai", GOOGLE_CLOUD_PROJECT, and GOOGLE_CLOUD_LOCATION
AI Gateway Set selectedType: "gateway"

Shared ACP broker

The plugin starts a shared ACP broker session so multiple commands reuse a single Gemini process. The broker starts on the first command that needs it and runs for the duration of the Claude Code session. If the broker is busy or unavailable, the plugin falls back to a direct Gemini CLI process automatically.

Execution modes

  • Default (--yolo --sandbox) -- Gemini can read and write files within the repository worktree, scoped by Gemini CLI's built-in sandbox mode (filesystem access limited to the working tree).
  • Plan (--plan) -- Read-only. Gemini proposes changes but does not execute them.

Session persistence and timeouts

Task sessions are indexed in the plugin's state directory (outside the working tree), enabling the --resume flag to continue from a prior thread (up to 20 recent sessions are retained).

The default task turn timeout is 30 minutes. Override with GEMINI_TASK_TIMEOUT_MS:

export GEMINI_TASK_TIMEOUT_MS=3600000  # 60 minutes

Attribution

Note

This plugin is derived from and inspired by OpenAI's Codex plugin for Claude Code, licensed under the Apache License 2.0. The broker lifecycle, state management, job control, rendering, and arg parsing modules are forked from that codebase with identifiers adapted for Google's Gemini CLI integration. See the NOTICE file for full attribution details.


FAQ

Do I need a separate Google account? You need Gemini CLI access. Configure your auth method in ~/.gemini/settings.json. The setup command checks this for you.

Does the plugin run a separate process? Yes. The plugin starts the Gemini CLI as a local process communicating via ACP (JSON-RPC 2.0 over stdio). A shared broker session keeps one Gemini process alive across multiple commands.

What is ACP? The Agent-Client Protocol is a JSON-RPC 2.0-based protocol used by the Gemini CLI for programmatic interaction. It supports session management, turn-based prompting, streaming updates, and session mode control.

Can I use this alongside the Codex plugin? Yes. Both plugins use separate namespaces (/gemini:* and /codex:*) and independent runtimes. They do not interfere with each other.

What happens if Gemini is not authenticated? The plugin detects this and tells you to run /gemini:setup. You can also run gemini interactively in a terminal to complete the OAuth flow.

How does the review gate work? When enabled via /gemini:setup --enable-review-gate, a Stop hook fires every time Claude is about to finish. The hook triggers a Gemini review of Claude's changes, giving you an automatic second opinion. Disable it with /gemini:setup --disable-review-gate.

Can I control which Gemini model is used? Yes. Pass --model <alias> to /gemini:rescue. See the model aliases table above. If no model is specified, Gemini uses its default.

Warning

The default --yolo --sandbox mode allows Gemini to modify files in your repository. Make sure your work is committed or stashed before running rescue commands that may write to the worktree.


License

Apache License 2.0. See LICENSE.