08 · 131 projects

Jev for safety & guardrails

Screening messages, tool calls and outputs before they cause trouble. Jev gives a fast, calibrated verdict that can sit in front of every request or agent action.

What Jev decides: Allow, flag or escalate a message, tool call or action.

Projects, newest first

Safety & guardrailsautomater.aiMCP tools the agent can skip are not a gate.Site · AutomaterSafety & guardrailsLangChain's Jev tool-call gateTool output stays out of the classifier's input.Site · Yurii OksamytnyiSafety & guardrailsjev-guard (klauswg)Real-time risk triage gateway for exchange deposits/withdrawals: TypeSafe Jev answers risk/pattern/freeze questions; Java code owns hard rules,…GitHub · ★ 35 · klauswgSafety & guardrailsjev-agent-authorizationKinde starter: identity + permissions, then TypeSafe Jev typed gate before every MCP tool call — allow, step-up approve, or stop.GitHub · ★ 5 · kinde-starter-kitsSafety & guardrailsjev-pilotFast System-1 Decision, Arbitration & Safety Engine for Autonomous AI Agents (Powered by TypeSafe Jev).GitHub · ★ 3 · h0j5bz0adh0-stackSafety & guardrailsA CAPTCHA rebuilt with JevThe browser measures how you fill the form.X post · Jarek CeborskiSafety & guardrailsAn agent-agnostic permission gateJev and hooks, so a dangerous command asks first.X post · ImRobotSafety & guardrailsjev-model-tokengateStreaming LLM proxy that holds tokens in a sliding buffer, runs semantic safety criteria (TypeSafe Jev by default) on each window, and releases or…GitHub · ★ 4 · Thanh-Mathieu95Safety & guardrailsA phishing message detectorTypeSafe Jev checks messages for phishing.X post · choroshinSafety & guardrailsEvery rule checked in 348 msJev reads each Claude Code reply and edit against the rules you wrote.X post · erkamyaman_ngSafety & guardrailsJev blocks a $50,000 transfer95% irreversible risk and 94% sensitivity, so it refused.X post · fluixooSafety & guardrailsJev finds the sensitive files in a shareA security researcher picks out risky files with Jev.X post · _xpn_Safety & guardrailsTesting Jev for prompt injectionA demo that shows how to test prompt injection against Jev.X post · Frank ChenSafety & guardrailsjev-chat-jarvis-macA macOS overlay that judges message intent and risk, then drafts.GitHub · ★ 346 · jev-chatSafety & guardrailsjev-edgeTyped-judgment admission control at the traffic edge: three-layer prompt-injection and abuse filter for nginx/OpenResty, powered by TypeSafe Jev.GitHub · ★ 34 · kiwi0719Safety & guardrailsjev-guard (muratcakmak)Claude Code plugin with PreToolUse-style hooks that combine local regex rules and batched TypeSafe Jev noul scores to deny rule-breaking edits and…GitHub · ★ 7 · muratcakmakSafety & guardrailsjevcCompile natural-language agent rules (and JSON Schema) into TypeSafe Jev programs: narrow typed questions plus a code-owned reducer/verdict you can…GitHub · ★ 6 · doronpSafety & guardrailsjev-harness (ismaelsoilet)Zero-dependency System One decision harness: 5 semantic gates saving frontier AI agent tokens on trivial errors & doom loops.GitHub · ★ 6 · ismaelsoiletSafety & guardrailsjev-idsBlazing-Fast Token-Efficient Intrusion Detection System (IDS) based on TypeSafe's Jev.GitHub · ★ 3 · jev-idsSafety & guardrailsclaude-jev-pluginTypeSafe Jev semantic guardrails for Claude Code.GitHub · ★ 2 · dr-dimitruSafety & guardrailstoxic-filterChosen expressions on X get blurred out.GitHub · eijiarakiSafety & guardrailstwitter-jev-guardLow-quality, spam and ad posts get a translucent watermark.GitHub · ★ 9 · qs-lllSafety & guardrailsdsh-jev-toolsJudgment, not generation, inside DeepSeek Harness.GitHub · ★ 5 · HorusJiangSafety & guardrailsagi-jev-containmentAGI JEV Detection — local AI agent monitor: chain-level malicious-agent detection (TypeSafe Jev + Sentinel), escalate-only L1–L5 containment, Neo4j…GitHub · ★ 4 · carlosedm10Safety & guardrailstweet-911Real-time AI / solicitation / parrot-bot scores for X posts and replies.GitHub · ★ 3 · elliothuxSafety & guardrailsjevshieldPython runtime security gate for agent tool calls: TypeSafe Jev Choice/Noul/Score dual-factor evaluation with fail-closed parsing,…GitHub · ★ 2 · lgy1027Safety & guardrailsjev-firewallReal-time firewall for AI coding agents: a PreToolUse hook checks every Claude Code / Codex tool call with deterministic shell-aware rules, and…GitHub · ★ 1 · Koushik890Safety & guardrailsA confidence badge on every postPosts on screen classified live as scam, slop or neither.X post · tanavtwtSafety & guardrailsJcyberA bug-bounty chain where Jev gates every step.GitHub · Jerry XiaoSafety & guardrailsslop-graderA CLI that audits text against your own rulesets.X post · VintuxaiSafety & guardrailsUXRayA screen overlay that flags manipulative interface patterns.GitHub · Mohammad ZohaibSafety & guardrailsagent-chaperoneScreen an agent's tool calls with Jev before they run.GitHub · ★ 20 · agent-chaperoneSafety & guardrailsjev-belayClaude Code Stop hook that blocks an unverified done: reads the transcript for evidence, asks Jev once, fails open on everything else.GitHub · ★ 18 · valentynkitSafety & guardrailspi-jev-sentinelPi coding-agent extension: TypeSafe Jev checks for tool calls, tool outputs and replies (prompt injection, approvals, secret scrubbing, task pinning).GitHub · ★ 9 · harshwasanSafety & guardrailsgg-friggin-ezFast, drop-in profanity and toxicity screener for Node.js, powered by TypeSafe AI Jev.GitHub · ★ 7 · ItisShikharSafety & guardrailsconstruct-auto-classifierEffect-based safety gate for AI coding agents' shell commands (OpenCode, Antigravity): fast structural rules, then TypeSafe's Jev or a chat model…GitHub · ★ 3 · godspedeSafety & guardrailsrh-guardReward-hack radar for coding agents: structural denies + TypeSafe Jev System One sidecar for Claude Code & Cursor hooks.GitHub · ★ 3 · 24601Safety & guardrailsJev-AVAntivirus scanner and real-time defense sentinel: extracts structural features from executables, scripts, and documents (entropy, hashes, strings, PE…GitHub · ★ 2 · newuser7171Safety & guardrailsjev-pii-checkerCLI that scans text and files for PII with TypeSafe Jev presence/sensitivity judgments plus regex and segmentation layers, emitting JSON or tables…GitHub · ★ 2 · coo-quackSafety & guardrailstoolgate (RiskAverseTech)Open tool-call firewall for AI agents: static rules first, then TypeSafe Jev risk judgments, shipping as a Claude Code PreToolUse hook and an MCP…GitHub · ★ 2 · RiskAverseTechSafety & guardrailsactiongate-jevRuntime authorization and guardrails for AI-agent tool calls with deterministic policy and TypeSafe Jev via OpenRouter.GitHub · ★ 2 · omkarghugarkar007Safety & guardrailsstepwardenEvery tool call your agent makes, checked before it runs.GitHub · ★ 1 · getexcitedSafety & guardrailsjev-crawlersBug-hunting crawlers built from five Unix-style primitives.GitHub · russfrankySafety & guardrailsjev-gatesSix hook gates for Claude Code that catch broken rules, out-of-scope edits, unfinished asks, and false claims, each judged by Jev against a fixed…GitHub · rashedInt32Safety & guardrailsjev_antispam_botA Telegram bot that deletes only the spam Jev is sure about.GitHub · ★ 11 · backmeupplzSafety & guardrailsmastra-jev-moderationInput moderation for Mastra agents in one file.GitHub · ★ 3 · CodeAlive-AISafety & guardrailsclear-headClaude Code Stop hook that checks factual claims in the assistant’s answer against what it actually read this session, using TypeSafe Jev as the…GitHub · ★ 2 · VladyslavHontarSafety & guardrailsjev-decisionsSafety checks for Hermes and other AI agents before they act, ask for approval, or verify a change.GitHub · ★ 2 · bojansandhausSafety & guardrailsBouncerA judgment layer that checks every Claude Code tool call against your policy.GitHub · ClownwareSafety & guardrailsfirehose-judgeThe live Bluesky firehose, judged post by post, with a lane for humans.GitHub · Leo MataSafety & guardrailsInterlockA gate on every agent tool call, with Jev as one of its sensors.GitHub · somooreSafety & guardrailsis-maliciousA codebase scanner that helps you not run malicious code.GitHub · ★ 23 · luantakSafety & guardrailshermes-jev-approvalsPoC: TypeSafe Jev as the reviewer for Hermes Agent smart command approvals.GitHub · ★ 14 · anpicassoSafety & guardrailsjevcalChoose acceptance thresholds from labeled examples.GitHub · ★ 11 · abhixhekSafety & guardrailspi-heedChecks every side-effecting tool call against what you asked for.GitHub · ★ 7 · NyarlathotepppppJSafety & guardrailsjevmodModeration with a probability per category and thresholds you set.GitHub · ohernandezdevSafety & guardrailsMade a real-time slop detector with jev as you scrollMade a real-time slop detector with jev as you scrollX post · RBilgilSafety & guardrailsPi WardenFlag risky actions and unverified completion.GitHub · ★ 143 · DevMortimerSafety & guardrailsJev Moderation BotScreen Discord messages and escalate concerns.GitHub · ★ 48 · brainstormitySafety & guardrailsjev-guardSecurity hook for coding agents: TypeSafe Jev risk-scores every tool call with session context (deny / ask / allow), flags prompt injection in…GitHub · ★ 32 · leepokaiSafety & guardrailszod-jevAdd meaning-based checks to shape validation.GitHub · ★ 9 · jomatsuSafety & guardrailsJev ShieldCheck MCP calls, outputs, and tool descriptions.GitHub · ★ 4 · caiovicentinoSafety & guardrailsJodCombine local schemas with typed model judgments.GitHub · ★ 4 · mateonunezSafety & guardrailspi-typesafe-jevA pi extension that exposes TypeSafe (Jev, System One) judgments as five pi tools, so a model can make narrow semantic judgments while your code and…GitHub · ★ 3 · legacybridge-techSafety & guardrailsjev-audio-beeperLow-latency audio censorship POC using Jev typed decisions and ffmpeg.GitHub · ★ 2 · santos-sanzSafety & guardrailstoolgateAgent tool/MCP call gate — allow / ask_human / deny via TypeSafe Jev.GitHub · ★ 2 · ndolinschiSafety & guardrailsAgent Handoff GateReview evidence before an agent hands work off.GitHub · ★ 1 · zsoXiSafety & guardrailsCheck RiskRecommend checks for a proposed code change.GitHub · ★ 1 · moezubairSafety & guardrailsSwitchboardExplore model routing alongside safety checks.GitHub · ★ 1 · aniruddh-krovvidiSafety & guardrailsguard-jevComment-moderation playground: paste a comment, Jev decides what to do with it.GitHub · ★ 1 · NorbertBodzionySafety & guardrailsjevegisGuardrails for LLM apps in one API call.GitHub · ★ 1 · 0xArxSafety & guardrailsJev as an agent safety monitorChecks each agent action first: most attacks caught, almost no false blocks.X post · NickSafety & guardrailsTypeSafe TypewriterSixteen judgments about your text, asked again on every keystroke.Site · Steve KrouseSafety & guardrailsAgentgateway exampleAdd a Jev check at an agent gateway.GitHub · agentgatewaySafety & guardrailscartshieldCartShield — SMB checkout fraud disposition via TypeSafe Jev.GitHub · ndolinschiSafety & guardrailsFraud detection with Jev and Kimi K3Jev sorts 100 emails in 1.42 s and sends the unsure ones to Kimi K3.X post · HassanSafety & guardrailsJev DetectorAn AI slop detector built on Jev, scanning about 10,000 words in 2 seconds.X post · Jozef GhermanSafety & guardrailsomp-jevens-classifierJev-powered model-judged permission gate for OMP (TypeSafe System One).GitHub · STRMLSafety & guardrailspkg-gatePre-install security gate for npm lifecycle scripts using TypeSafe System One.GitHub · hemanthSafety & guardrailsscam-shieldScam text message filter powered by TypeSafe's Jev model.GitHub · ShupingRSafety & guardrailsspendbrakeAgent budget brake — continue / downgrade_model / stop via TypeSafe Jev.GitHub · ndolinschiSafety & guardrailstrustgateTrustGate — indie media T&S gate via TypeSafe Jev.GitHub · ndolinschiSafety & guardrailsJev for security engineeringKostas on threat hunting with Jev: define the questions and outputs, then rank, score and classify activity at scale.X postVSafety & guardrailsVercel fx: Jev as a command safety reviewerGuillermo Rauch: Jev reviews every fx command, faster and more accurate than a chat model.X postSafety & guardrailspi-jev-sentinelAdd TypeSafe Jev checks for coding-agent tool calls, tool outputs, and replies—shipping as a Pi extension plus a shared Claude Code / Codex CLI hook.GitHub · harshwasanSafety & guardrailsexample.com** Jev CLI is out** <a:elmo_fire:1549940287235686410> `jev` puts TypeSafe's Jev model on the command line.Site · NasrSafety & guardrailsagy-jevgateFail-closed Antigravity (agy) PreToolUse hook for run_command: exact read-only fast-pass, a short static destructive guard, then TypeSafe Jev risk…GitHub · catpotdSafety & guardrailsbeam-cli (AgentBeam)Local AgentBeam CLI (beam): installs native hooks for detected coding agents, captures agent/MCP activity, applies policy locally, and optionally…GitHub · whyashthakkerSafety & guardrailsproject_blackoutBuilt a simulated SIEM to showcase Jev's risk assessment capabilities.GitHub · DataSherpaSafety & guardrailsclaude-code-jevExperimental Claude Code PreToolUse permission gate that classifies each tool call as allow / block / ask with TypeSafe Jev via OpenRouter’s typed…GitHub · RahulBalakaviSafety & guardrailsdsh-jev-interceptorDeepSeek Harness plugin: TypeSafe Jev risk-classifies tool calls (fail-closed / shadow→enforce) and can score session-reference messages instead of…GitHub · AskTheWaySafety & guardrailsgrok-jev-guardTyped preflight/approval layer for Grok Bot: local policy owns hard boundaries; TypeSafe Jev judges ambiguity; the agent executes only inside the…GitHub · 0xwhrariSafety & guardrailsHeimelHEIMEL solves a simple problem: agents can decide and plan, but something still has to control what actually happens.GitHub · NjaalSafety & guardrailsdigital-twinsHey! For the past few months I've been building simulated sandboxes as an approach to buid safer ai agents.GitHub · hadesSafety & guardrailslesswrong.comHi everyone, I just wrote up a small experiment testing TypeSafe’s Jev as a trusted and cheaper monitor alternate for AI Control.Site · venkatSafety & guardrailsJeevesI added twitch moderation to so you can set actions like ``` !setaction ban users if they harass others !setaction time a user out for 1 week if they…GitHub · JustSomeDevSafety & guardrailsjev-guardI built "jev-guard", an auto-approval layer for Claude Code / Codex / Antigravity agent harnessesGitHub · sudo rm -rf /*Safety & guardrailspi.devim using jev as a fast judgement layer for the agent; checks each tool call (irreversible?Site · devmortimer24Safety & guardrailsJev as a terminal guardrail.Jev as a terminal guardrail.Site · lampSafety & guardrailsJev Call ScreenerOpen-source call-screening backend: ask why the caller is calling, classify the transcript with TypeSafe Jev, and forward or reject under a fail-open…GitHub · SuchintKSafety & guardrailsJev Content GuardA Manifest V3 Chrome extension that filters fraud, advertising, AI slop, spam, clickbait, "info-gypsy" schemes, and toxicity from web pages using…Site · serejkaaa512Safety & guardrailsjev-cloud-cost-guardianGitHub Action FinOps gate: normalize cloud cost evidence, ask TypeSafe Jev for a typed approve / warn / block / manual-review decision, then apply…GitHub · JevForgeSafety & guardrailsjev-cmdline-classifierPortable agent skill plus stdlib Python/JS reference that classifies shell commands with TypeSafe Jev Choice into allow / prompt / forbidden for…GitHub · liodaliSafety & guardrailsjev-fuseGoverned reverse proxy between callers (Claude Code, MCP, SDKs) and TypeSafe Jev or local Laya: turn probabilities into ALLOW/ASK/DENY-style actions…GitHub · 0xshikharSafety & guardrailsjev-gatesComposable three-valued semantic logic circuits: small TypeSafe Jev (or LocalJev) judgments become explicit TRUE/FALSE/UNKNOWN signals, combined with…GitHub · carlchou0dailyfreshSafety & guardrailsjev-guard (CMaintz)Framework-agnostic TypeScript guardrail that vets proposed agent tool calls with TypeSafe Jev (score/noul dimensions → allow/block/hold), plus…GitHub · CMaintzSafety & guardrailsjev-prompt-sentryIngress reverse proxy for Anthropic Messages: one batched TypeSafe Jev call screens jailbreaks, indirect injections, and exfil risk before the…GitHub · ca7aiSafety & guardrailsjev-sap-commerceSAP Commerce (jevintegration) extension: TypeSafe Jev moderates product reviews and suggests categories with dry runs, audit records, and human…GitHub · EmenowiczSafety & guardrailsjev-x-filterChrome MV3 extension that filters X (Twitter) timeline spam with TypeSafe Jev typed decisions across six toggleable classes; high confidence only…GitHub · harodgggSafety & guardrailsJev_validation_agentPython Jev Guard (jev_guard): double-check LLM answers with TypeSafe Jev Noul/Score/Choice checks (on-topic, contradicts-source, format) before…GitHub · omkarchougule19Safety & guardrailsjevgate (craxrev)Claude Code plugin: before Bash/Write, TypeSafe Jev reports risk facts; fixed rules allow, ask, or deny—aimed at bypassPermissions mode.GitHub · craxrevSafety & guardrailsJevGuardClaude Code / Codex plugin that turns project rules (from CLAUDE.md / AGENTS.md) into typed Jev checks on PreToolUse and Stop hooks, so the agent…GitHub · Jhonnyr97Safety & guardrailsJevGuard (blacksinisterx)Agent tool-execution security layer: hard-rule prefilter, then TypeSafe Jev allow/review/block (distinct from catalog leepokai/CMaintz/klauswg…GitHub · blacksinisterxSafety & guardrailsJevSlopHosted and source-built app that scores note articles for AI-slop writing patterns with TypeSafe Jev (multi-axis Score plus overall Choice).Site · TKY-27Safety & guardrailsJuardrailsGo guardrails management service for TypeSafe Jev: define Choice/Score/Noul policies in YAML or a UI, batch questions into one provider call, then…GitHub · abhaybhargavSafety & guardrailsmayiPermission gate for Claude Code, Cursor, and Codex: TypeSafe Jev scores each shell/file/MCP tool call before it runs; routine work passes silently…GitHub · AidenHadisiSafety & guardrailsModels are getting smarter.Models are getting smarter.GitHub · ThibautSafety & guardrailsOpenCode Security GuardLinux OpenCode shell permission guard: local command check plus TypeSafe Jev read-only score ≥ 0.90 for auto-allow (ask otherwise; deny stays final).GitHub · koppertSafety & guardrailsopencode-jev-guardOpenCode 2 plugin: every local shell (and FarHand remote shell) command is judged by TypeSafe Jev before run—overall verdict plus host-litter /…GitHub · CogFluxSafety & guardrailspi-jev-permitPi coding-agent extension that asks TypeSafe Jev whether each bash / write / edit call should run, after local hard-deny, allow/deny rules, and a…GitHub · kurihadaSafety & guardrailspi-typesafe-bash-guardPi coding-agent extension that classifies every bash tool call and user !GitHub · gowthamgtsSafety & guardrailsPrompt RejectorMCP server and local HTTPS API that screen prompts, skill files, and MCP tool descriptions before an agent acts: deterministic checks plus TypeSafe…GitHub · revsmokeSafety & guardrailsRegret CheckChrome extension that pauses you before commit-like clicks you might regret: a local server scores the page action with TypeSafe Jev (via OpenRouter)…GitHub · daniloceciliaSafety & guardrailsResponsible AI HarnessModel-agnostic assessment harness: deterministic hard rules plus an optional TypeSafe Jev judge for prompt injection, secret/PII leakage, unsafe tool…GitHub · syabdulrSafety & guardrailsagent-control-planeSharing what I'm building: ZIFFER, a deterministic authorization layer for AI agents.GitHub · Yacine "ZIFFER"Safety & guardrailsSoterAutomated Discord moderation: TypeSafe Jev (OpenRouter Decisions) scores hate speech and spam per message; code deletes clear hits, flags borderline…Site · frolleksSafety & guardrailssecond-thoughtterminal seatbelt on Jev — judges shell commands before Enter, ~$0.000001/check, blocks past 85% confidenceGitHub · rodriveigaSafety & guardrailsTripwireOpenAI-compatible streaming proxy that watches partial LLM completions and can abort the upstream generation mid-flight.GitHub · anuran-deSafety & guardrailsWatermelonStatus-update honesty auditor: paste a weekly program update; TypeSafe Jev judges language while local code parses dates/slippage; the app returns…Site · shashwatc12Safety & guardrailsLLM guardrailsEvaluate incoming and outgoing messages.GuideSafety & guardrailsCitation checksTest a claim against a supplied source passage.Guide

Other use cases

Agents & browsers 193Coding & code review 227Routing & model choice 161Context & memory 66Search & RAG 56Data & extraction 61Documents & research 28Support & inbox 45Marketing, sales & social 95Trading & finance 46Games & real time 173Creative & generative UI 40Home, robots & devices 20Everyday apps 113SDKs & integrations 308Benchmarks & research 332Guides & docs 202