frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Vigil – Zero-dependency safety guardrails for AI agent tool calls

2•HexitLabs•2h ago
We run 15 AI agents on a production server with full shell access. One of them tried to rm -rf a directory it shouldn't have touched. Another started curling cloud metadata endpoints. We wrote some hardcoded rules to catch the obvious stuff, then realized we were building the same safety layer everyone else will need too. So we extracted it into a library.

Vigil is a deterministic rule engine that inspects AI agent tool calls before they execute. 22 rules across 8 threat categories: destructive shell commands, SSRF, path traversal, SQL injection, data exfiltration, prompt injection, encoded payloads, and credential exposure. It's not an LLM wrapper — we don't trust an LLM to guard another LLM. Pure pattern matching, zero dependencies, <2ms per check, works completely offline.

npm install vigil-agent-safety

import { checkAction } from 'vigil-agent-safety'; const result = checkAction({ agent: 'my-agent', tool: 'exec', params: { command: 'rm -rf /' }, }); // result.decision → "BLOCK" // result.reason → "Destructive command pattern" // result.latencyMs → 0.3

It plugs into MCP servers, LangChain tool chains, Express middleware, or anything else. MIT licensed, no API keys, no network calls, no telemetry.

This is v0.1 — probably too aggressive for some use cases. Next up is a YAML policy engine (v0.2) and an MCP proxy. We'd love feedback on the rule set, false positive experiences, and threat categories we're missing.

GitHub: https://github.com/hexitlabs/vigil

Comments

HexitLabs•2h ago
Author here, happy to answer any questions.

Some context on why we built this: you might have seen the post earlier this week about someone building a file recovery tool after Claude Code rm -rf'd their Obsidian vault through a symlink. We had similar near-misses running our own agent swarm, agents curling cloud metadata endpoints, attempting path traversal, executing destructive commands during "cleanup" steps. We kept adding one-off guards and eventually realized this should be a proper library.

The main design choice was making it deterministic rather than using an LLM to review tool calls. An LLM guarding another LLM felt like asking the fox to guard the henhouse. Pattern matching is boring, but it's fast, predictable, and works offline.

Happy to hear about false positives, missing threat categories, or use cases where the rule set is too aggressive. That's the main thing we want to calibrate for v0.2.

Show HN: A VCluster in Docker with Terraform and Istio

https://github.com/madduci/vind-with-mesh
1•madduci•2m ago•0 comments

Show HN: Colored Title Bar – unique colors per VS Code workspace

https://github.com/wstuckey/colored-title-bar
2•wstuckey•4m ago•0 comments

Echoslate – Offline todo and Kanban tracker for programmers (MIT, .NET 8)

https://github.com/pjmavcom/echoslate
1•pjmavcom•4m ago•0 comments

Show HN: Bridge your Claude/OpenAI subs into a team API with per-key cost caps

https://github.com/Shreyas-Dayal/ai-cli-bridge
1•shreyas8•4m ago•0 comments

FounderSpace – AI-guided startup validation in 30 minutes

https://www.founderspace.work
1•VladCovaci•7m ago•1 comments

P5.lcd

https://p5.lcd.tulv.in/
1•atulvi•12m ago•0 comments

US and Israel carrying out strikes against Iran

https://www.cnn.com/world/live-news/israel-iran-attack-02-28-26-hnk-intl
16•simonebrunozzi•15m ago•3 comments

The March of Nines

https://kk.org/thetechnium/the-march-of-nines/
1•ahmaman•16m ago•0 comments

Song about afterlife from AI perspective [video]

https://www.youtube.com/shorts/DZGvrqNP7lg
1•modinfo•16m ago•0 comments

Show HN: TapPause – A dead-simple timer for intentional breaks

https://apps.apple.com/hk/app/tappause-pause-timer/id6758304453
1•louischen•17m ago•1 comments

I built a startup game because I wanted one to play

https://capturn.io
3•disobae•17m ago•1 comments

U.S. and Israel Conduct Strikes on Iran

https://www.nytimes.com/live/2026/02/28/world/iran-strikes-trump
14•gammarator•17m ago•0 comments

Japan aims for world first in space-based solar power

https://www.asahi.com/ajw/articles/16293144
2•helsinkiandrew•19m ago•0 comments

Simulation for Agentic Evaluation

https://yortuc.com/posts/simulation-for-agentic-evaluation/
2•xyortuc•21m ago•0 comments

Israel says it has attacked Iran, declares state of emergency

https://www.cnbc.com/2026/02/28/trump-us-military-iran-strikes-middle-east-oil.html
17•LordAtlas•23m ago•2 comments

Show HN: PicShift – Convert images in the browser using WebAssembly

https://picshift.app/
1•pod4g•24m ago•1 comments

I Just Cancelled My ChatGPT Pro Plan

https://aifordevelopers.substack.com/p/i-just-cancelled-my-chatgpt-pro-plan
3•dariubs•28m ago•0 comments

Trump directs US agencies to toss Anthropic's AI

https://www.reuters.com/world/us/trump-says-he-is-directing-federal-agencies-cease-use-anthropic-...
2•frb•34m ago•0 comments

Huk vs. Naniwa Starcraft [video]

https://www.youtube.com/watch?v=rFllSVjK5hY
1•marysminefnuf•36m ago•0 comments

Show HN: I built GeoQuests where people can request photos of a place

https://geoquests.io
2•Swain123•38m ago•0 comments

Israel launches attack on Iran, defence minister says

https://www.bbc.com/news/live/cn5ge95q6y7t
14•Philpax•41m ago•4 comments

Israel launches strike against Iran, declares state of emergency across country

https://www.cnn.com/2026/02/28/middleeast/israel-attack-iran-intl-hnk
24•lavp•41m ago•6 comments

Ask HN: AI driven "legacy" Debug and Monitoring Tools

1•chris_money202•44m ago•0 comments

AI Mistakes Are Infuriating Gamers as Developers Seek Savings

https://www.bloomberg.com/news/articles/2026-02-25/ai-mistakes-are-infuriating-gamers-as-develope...
4•breve•45m ago•0 comments

AI Doomer Doublethink: The Orwellian Worldviews in the AGI Discourse

https://internetofbugs.substack.com/p/ai-doomer-doublethink-the-orwellian
2•edward28•45m ago•0 comments

EUrouter – Integrate the latest AI models, without sending data outside the EU

https://www.eurouter.ai
4•fahrradflucht•47m ago•0 comments

What's cooking in git.git (Feb 2026, #11)

https://lore.kernel.org/git/xmqq8qcdof3f.fsf@gitster.g/
3•todsacerdoti•51m ago•0 comments

Poker4U – Educational game theory and better life decisions

https://github.com/Lameda12/poker4u
4•Alamedin_ishere•53m ago•1 comments

Varlock: .env Files Built for Sharing

https://github.com/dmno-dev/varlock
2•appwiz•54m ago•0 comments

Saving No One

https://netwars.pelicancrossing.net/2026/02/27/saving-no-one/
2•ColinWright•57m ago•0 comments