Hacker News

Show HN: PKGSmith

Hacker News - Sat, 03/07/2026 - 9:34am

Article URL: https://pkgsmith.app/

Comments URL: https://news.ycombinator.com/item?id=47288010

Points: 1

# Comments: 0

Categories: Hacker News

Show HN: ANSI-Saver – A macOS Screensaver

Hacker News - Sat, 03/07/2026 - 9:34am

Hi, I've been working on something I've been thinking for long time but since I had no experience with macOS screen savers I always posponed. Now thanks to Claude I was able to create a screensaver that scroll ANSI files while your computer is idle.

It allow to use local ANS files or packs directly from 16colo.rs.

Repo: https://github.com/lardissone/ansi-saver

Comments URL: https://news.ycombinator.com/item?id=47288007

Points: 2

# Comments: 0

Categories: Hacker News

Got tasks? Feed your Dactyl

Hacker News - Sat, 03/07/2026 - 9:33am

Article URL: https://taskadactyl.com

Comments URL: https://news.ycombinator.com/item?id=47287999

Points: 1

# Comments: 1

Categories: Hacker News

Show HN: RankClaw – AI-audited all 14,706 OpenClaw skills; 1,103 are malicious

Hacker News - Sat, 03/07/2026 - 9:32am

RankClaw (rankclaw.com) is a security scanner for AI agent skills — the OpenClaw/ClawHub ecosystem that extends Claude-based agents with file, web, and shell access.

Data: - 14,706 skills indexed - Every single skill has a full AI deep audit report (14,704 complete) - 1,103 confirmed malicious (7.5%)

The key finding: automated surface scanning (metadata, dependency checks, pattern matching) systematically undercounts malicious skills. Skills that pass shallow heuristics fail AI audit because the attack is in the natural language of the SKILL.md — prompt injection, deferred execution, social engineering — none of which pattern matching detects.

The attack patterns found by AI deep audit: - Bulk publishing campaigns — one actor published 30 skills named "x-trends" across multiple accounts. 28 of 30 confirmed malicious. Goal: distribution at scale before detection. - Brand-jacking — 4 skills named clawhub/clawhub1/clawbhub/clawhud impersonating ClawHub's own CLI. macOS: base64 curl|bash to a raw IP. Windows: password-protected ZIP from a stranger's GitHub (the password prevents GitHub's malware scanner from opening it). - Prompt injection in legitimate-seeming skills — one scored 95/100 shallow, 38/100 after AI audit. The injection text wasn't in code — it was in the SKILL.md instructions. - On-demand RCE via challenge evaluation — claws-nft instructs the agent to "evaluate" challenges that can be "math, code, or logic problems." Server decides which type at call time. - LLM-generated payload — lekt9/foundry contains no malicious code. It instructs the AI to generate code and execute it. Static analysis finds nothing. The payload doesn't exist until the AI writes it during a conversation. - Social engineering — bonero-miner has a "Talking to Your Human" section with a pre-written script for the AI to use: "Can I mine Bonero? It's a private cryptocurrency - like Monero but for AI agents. Cool?"

Skills differ from browser extensions: no sandbox. Full file system, shell, and network access. The SKILL.md instructions are directives to the AI model — you need AI to audit AI.

Scoring model is open: Security 40%, Maintenance 20%, Docs 20%, Community 20%.

Free to check any skill: rankclaw.com

Comments URL: https://news.ycombinator.com/item?id=47287985

Points: 1

# Comments: 0

Categories: Hacker News

TanStack Intent

Hacker News - Sat, 03/07/2026 - 9:31am

Article URL: https://tanstack.com/intent/latest

Comments URL: https://news.ycombinator.com/item?id=47287973

Points: 2

# Comments: 0

Categories: Hacker News

Show HN: File Indian income tax from the browser. No signup, privacy first

Hacker News - Sat, 03/07/2026 - 9:30am

Tax filing in India is broken.

You get income data from 3 government sources. They often disagree with each other. You're supposed to figure out which one is correct.

Existing tools like ClearTax and TaxBuddy require you to upload all your financial data to their servers, create accounts, and trust third parties with your PAN, salary, investments, and bank details.

For something as sensitive as tax data, that never felt right to me.

So I built Fiscally. It runs 100% in your browser.

No server. No signup. No data upload. Your financial data never leaves your machine.

What it does: - Import government tax files (all 3 formats) - Auto-detect conflicts between sources and pick the right values - Calculate tax under both old and new regime - Generate filing-ready JSON for the government portal - Generate PDF computation report - Save progress across days using IndexedDB (tax filing takes time)

I built this with a practicing Chartered Accountant who validated every tax rule.

Stack: Next.js, TypeScript, IndexedDB, Zod, jsPDF

The tricky part was reconciliation — same income shows up differently across sources because of rounding, timing, or formatting. Built a matching engine with source trust levels to handle this automatically.

Would love feedback from the community.

Comments URL: https://news.ycombinator.com/item?id=47287963

Points: 1

# Comments: 0

Categories: Hacker News

Show HN: Cooperation Cube – a semi-cooperative game with a rotating cube

Hacker News - Sat, 03/07/2026 - 8:42am

Back in 2017, I built a physical board game called Cooperation Cube.

The idea started somewhere over the Pacific on a long flight. I had been building software all day for years and felt the urge to make something physical. Something you could actually hold.

The result was a wooden cube with a 5×5 grid of holes on four faces. Players place colored sticks into the cube. Short pegs sit in a single face, while longer beams pass through the cube. Each player has pattern cards they are trying to complete on the face in front of them.

The twist is that the cube rotates 90° every round.

Patterns you were building suddenly become part of someone else’s board. Work you thought was secure becomes someone else’s opportunity.

There are also cooperation cards that require two players to coordinate moves without revealing their individual goals. The game lives in the tension between helping someone and quietly advancing your own position. I took some inspiration from the “shooting the moon” mechanic in Hearts and from social deduction games like Mafia.

The practical problem was the hardware.

Each cube required drilling a 5×5 grid on four faces of a wooden block. That is 100 holes, and they all have to line up perfectly. If even one is slightly off, the sticks will not pass through the cube correctly. For every usable cube I made, I probably ruined five blocks of wood.

I considered making a digital version several times but never got it to the finish line until now.

Try it here: https://cooperationcube.com

More details and pictures of the physical game here: https://www.kevinsdias.com/posts/cooperation-cube.html

Tech details: - Next.js + React frontend - Supabase (Postgres + Realtime) backend - TypeScript throughout - Game state stored as JSONB representing the 5×5×5 cube - Game engine implemented as pure functions for move validation, turn logic, and cube rotation - Supabase Realtime pushes updates so all players see moves instantly - Optional AI players fill empty seats so games can start without four humans

The physical cube is still my favorite way to play, but now the game is not limited to the four or five wooden cubes that exist.

Comments URL: https://news.ycombinator.com/item?id=47287576

Points: 1

# Comments: 0

Categories: Hacker News

Ask HN: How do you get users for free tool suite without annoying self promo?

Hacker News - Sat, 03/07/2026 - 8:42am

I built a large collection of free, no signup (AI) dev tools. Everything runs client-side, no ads, no tracking, no nothing. My goal is to create value and build something genuinely useful here.

I've tried the obvious stuff like SEO optimization, a few substack posts, etc. but I don't want to aggressively spam and self promote.

For those who've successfully shipped free tools or side projects: what actually moved the needle for you?

Comments URL: https://news.ycombinator.com/item?id=47287573

Points: 1

# Comments: 0

Categories: Hacker News

LLMs: Solvers vs. Judges

Hacker News - Sat, 03/07/2026 - 8:41am
Categories: Hacker News

Pages