Hacker News

I wanted Claude Code to notify my ESP32 when it finished a task. A `curl` in the Stop hook worked — until I needed a local dashboard too. The dashboard couldn't expose an HTTP endpoint, it needed to subscribe. And every new client meant editing hook configs across all my projects.

So I built agent-pulse: a local gateway that captures AI agent lifecycle events and fans them out to registered clients.

You register clients once in a global YAML. Projects just call `agent-pulse hook --provider claude --event stop` from their hooks. The gateway handles routing, per-client filtering, and parallel dispatch. It auto-starts on the first event.

Two delivery modes: HTTP POST fan-out for webhooks and scripts, SSE stream (`GET /events/stream`) for dashboards and anything that can't expose an endpoint. You can also attach project metadata via `.agent-pulse.json` that travels with every event.

Single Go binary, zero runtime dependencies. Claude Code and Gemini CLI supported today, OpenCode and Codex CLI in progress.

Note: Claude Code recently added native HTTP hooks, so you can POST directly to an endpoint without any intermediary. agent-pulse is useful when you have multiple clients, need SSE, or want one config across projects.

Thanks!

https://github.com/SantiagoBobrik/agent-pulse

Comments URL: https://news.ycombinator.com/item?id=47270638

Points: 1

# Comments: 0

Netwall (https://netwall.org) is an ultra-simple, text-only public wall.

- No accounts, no sign-ups.

- You type a message, hit enter, and it deletes itself after exactly 1 hour.

- If the community likes a post, they can click "+5m" to vote to keep it alive a bit longer.

Stack: Vanilla JS, Node/Express, Postgres. It features a few classic terminal themes (including Solarized Dark and the Openwall mailing list aesthetic), and sits behind OpenAI's Moderation API to keep it from turning into a complete swamp.

Let me know what you think!

Comments URL: https://news.ycombinator.com/item?id=47270637

Points: 1

# Comments: 0

This is a realtime collaborative blog! It works like git!

Comments URL: https://news.ycombinator.com/item?id=47270630

Points: 1

# Comments: 1

Hi HN,

I’m a dancer and small business owner who runs a Pilates studio in NYC. Recently I decided to do something slightly irrational and produce an evening-length dance theater piece independently.

Normally projects like this rely on grants or institutions, but I’ve been curious whether it’s possible to approach funding a piece of art more like a startup experiment.

The “product” is the performance itself. The costs are things like rehearsal space, paying the second performer, costumes, and building a small set.

Instead of waiting on grants, I started trying to raise the budget directly from people who believe in independent work.

It’s an unusual experiment and I’m learning a lot about what it takes to make something exist without institutional backing.

If anyone here has thoughts about creative funding models or ways artists might approach this differently, I’d genuinely love to hear them.

And if you’re curious about the project itself, I’m happy to share more.

Thanks for reading.

Comments URL: https://news.ycombinator.com/item?id=47270626

Points: 1

# Comments: 0

Article URL: https://mindgard.ai/blog/doctronic-is-now-accepting-new-patients-and-unsafe-instructions

Comments URL: https://news.ycombinator.com/item?id=47270614

Points: 1

# Comments: 0

Article URL: https://alexanderkustov.substack.com/p/academics-need-to-wake-up-on-ai-part

Comments URL: https://news.ycombinator.com/item?id=47270610

Points: 1

# Comments: 0

Article URL: https://meta-trials.vercel.app/

Comments URL: https://news.ycombinator.com/item?id=47270607

Points: 2

# Comments: 0

Article URL: https://twitter.com/levie/status/2013018817610518642

Comments URL: https://news.ycombinator.com/item?id=47270605

Points: 1

# Comments: 0

Job tailoring advice is everywhere but it all bottlenecks at the same place: actually doing it 50 times. Copy resume, paste into ChatGPT, rewrite bullets, reformat everything that broke, repeat. Ajusta sits on the job page (LinkedIn, Indeed, etc.), reads the posting, and rewrites your resume against the JD without touching your formatting or adding anything that isn't already there. The output is your resume with tighter language and better keyword alignment, not a new document built from a generic template. The constraint we cared about was hallucination. A lot of resume tools will invent responsibilities or rephrase something so aggressively it no longer reflects what you did. Ajusta only works with what's already in your resume. Still early. Would genuinely like feedback from anyone who tries it, especially on edge cases like non-standard formats or roles with heavy technical terminology.

Comments URL: https://news.ycombinator.com/item?id=47270574

Points: 1

# Comments: 0

Article URL: https://en.wikipedia.org/wiki/UzAuto_Motors

Comments URL: https://news.ycombinator.com/item?id=47270259

Points: 1

# Comments: 0

Article URL: https://theconversation.com/persian-gulf-desalination-plants-could-become-military-targets-in-regional-war-277597

Comments URL: https://news.ycombinator.com/item?id=47270245

Points: 1

# Comments: 0

Article URL: https://medicinalseedkit.com/kit/#aff=hfc13x9e67

Comments URL: https://news.ycombinator.com/item?id=47270243

Points: 1

# Comments: 0

Article URL: https://salesenablement.wordpress.com/2026/03/06/20-years-later-the-history-of-bizsphere-a-sales-enablement-martech-player-ahead-of-its-time/

Comments URL: https://news.ycombinator.com/item?id=47270237

Points: 1

# Comments: 0

I build SaaS products and WordPress plugins. One recurring problem kept coming up.

Feature requests everywhere.

Twitter replies. Emails. Support tickets. Random Slack messages. DMs from users.

The problem is not collecting feature requests. The problem is deciding what actually matters.

Votes alone don't solve that. The loudest users win.

So I built BuildVote.

It is a public roadmap + feature voting system, but with an intelligence layer that tries to understand the context behind feedback (yes, it is a constant improvement of teaching the system).

Instead of only counting votes, the system analyzes comments and feedback signals to help identify patterns and priorities.

The goal is simple: Help product builders prioritize features with better signal, not just noise.

Current features: - Public roadmap with feature voting - Context-aware feedback collection - AI-assisted prioritization insights - Changelog and roadmap transparency - Simple embed / hosted roadmap

The target users are: - Indie SaaS founders - WordPress plugin developers - small product teams

I built it because I wanted something I would actually use for my own products.

Still early. I'm still working on many things. In fact our product roadmap is the show-and-tell of the product.

I would love feedback from builders here.

https://buildvote.com

Comments URL: https://news.ycombinator.com/item?id=47270230

Points: 1

# Comments: 0

Article URL: https://github.com/vstorm-co/full-stack-ai-agent-template

Comments URL: https://news.ycombinator.com/item?id=47270227

Points: 1

# Comments: 0

Article URL: https://blog.yoshuawuyts.com/a-grand-vision-for-rust/

Comments URL: https://news.ycombinator.com/item?id=47270214

Points: 2

# Comments: 0

Article URL: https://www.youtube.com/playlist?list=PL1TfIrAqsOzcNSHV00pmvNusNTqYagQfz

Comments URL: https://news.ycombinator.com/item?id=47270212

Points: 2

# Comments: 0

Hi Hacker News! We're launching Zalor, an agent testing platform.

Agents often break when you tweak system prompts, swap models, or add tools. Zalor automatically generates test scenarios and evaluates your agent so you know it's reliable before deploying to production.

We currently support the OpenAI Agents SDK and are onboarding other frameworks. A GitHub integration is coming so you can get feedback on every update.

Looking forward to hearing feedback from people building agents.

Comments URL: https://news.ycombinator.com/item?id=47270208

Points: 3

# Comments: 3

Article URL: https://news.bloomberglaw.com/artificial-intelligence/netflix-acquires-ben-affleck-founded-ai-moviemaking-business

Comments URL: https://news.ycombinator.com/item?id=47270200

Points: 1

# Comments: 0

I built aidevshield, a free open-source security scanner that catches vulnerabilities in AI coding tool configurations things like Cursor, Copilot, and Cline setups.

The problem: AI coding assistants are becoming attack surfaces. In the past year we've seen real exploits:

Clinejection (Dec 2025): Prompt injection via GitHub issue titles → AI triage bot executes shell commands → malicious npm package published to 5M users tj-actions/changed-files (Mar 2025): Tag repointing attack hit 23,000 repos via pull_request_target Shai-Hulud (2025): First self-propagating npm worm using postinstall hooks Rules File Backdoor (Feb 2025): Hidden Unicode characters in .cursorrules for invisible prompt injection Existing tools cover parts of this — Zizmor and actionlint for GitHub Actions, Socket.dev for npm — but nothing scans across all four domains: workflows + AI configs + npm packages + prompt injection.

aidevshield does. One command, zero dependencies beyond js-yaml, fully offline:

npx aidevshield scan . What it detects:

Dangerous GitHub Actions patterns (wildcard users, untrusted event field interpolation, unpinned third-party actions) pull_request_target with untrusted checkout (Pwn Request pattern) Malicious npm lifecycle scripts (postinstall curl | sh, node -e inline exec) Hidden Unicode in AI config files (.cursorrules, .github/copilot-instructions.md) Wildcard tool permissions (Bash(), Edit()) Exposed .env files without .gitignore protection Cache poisoning + credential exposure combos Outputs text, JSON, or SARIF (for GitHub Code Scanning integration).

56 end-to-end tests. Every detection maps to a documented real-world attack.

MIT licensed. No signup, no paywall.

GitHub: https://github.com/aidevshield/aidevshield

Happy to answer questions about the threat landscape or implementation.

Comments URL: https://news.ycombinator.com/item?id=47270193

Points: 1

# Comments: 0