Agent tooling
MCP servers, model routers, and approval gates for people running coding agents on TypeSafe Jev.
This hub is for teams on Claude Code, Pi, Hermes, or a custom agent who need a fast route, allow, or compact decision on the hot path. Each tool sends Choice, Noul, or Score questions to Jev instead of parsing yes or no out of a chat reply. Prefer READMEs that state a latency budget and fail closed when the API errors.
Who it is for: Teams running Claude Code, Pi, Hermes, or custom agents who need fast, thresholdable decisions on the hot path instead of parsing YES/NO from chat completions.
How to choose: Start with the integration surface you already use (MCP vs npm plugin vs HTTP sidecar). Prefer repos that document latency budgets, failure modes, and which Jev primitives they send (Choice for routing, Noul for allow/deny, Score for severity). If the tool auto-approves tool calls, read how it fails closed when the API errors.
122 listings in this bucket · Learn Jev
See Jev alternatives compared, Kev vs Jev and Mapika Decider vs Jev.
Guide
What counts as agent tooling here
This hub lists routers, MCP servers, compaction helpers, review gates, and guardrails where System One sits between your agent and side effects. The pattern is the same: serialize the risky moment (shell command, tool call, diff hunk) into structured state, ask a typed question, threshold probabilities in code. Chat models may still write code or prose elsewhere; Jev is the judge on the narrow decision.
Primitives you will see most often
Choice picks one label from a small catalog (model name, browser action, approval verdict). Noul answers yes/no with a calibrated probability, perfect for allow/block gates. Score ranks ordered levels (blast radius, severity, compaction keep/drop). Many tools fan out several questions in one API call, then compose answers with policy text you control.
Highlights
Hand-picked listings with extra detail on this directory. Star counts and demos still live on each card.
- hermes-jev-approvals
Hermes smart approvals via typed Choice instead of a one-token chat completion. Strong fit if you search for hermes jev or jev hermes.
- typesafe-mcp
Single-binary MCP server and Go CLI (typesafe-mcp) for evaluate setup across Claude Code, Desktop, Codex, and Pi.
- pi-jev-auto-mode
Pi plugin that auto-approves safe bash/write/edit calls using Jev thresholds (pi-jev / pi jev queries).
- pi-warden
Pi supervisor that steers the agent back on track with Jev judging rules, risky tools, and stuck loops.
- jev-shield
Semantic MCP firewall: calibrated screening on tool calls, results, and poisoned descriptions.
- fast-jev-compaction
Compaction gate: Noul on whether a tool result deserves to stay verbatim in context.
FAQ
- What is Hermes Jev?
Short answer: see the Hermes Agent + Jev guide for the full setup, or the hermes-jev-approvals listing for just the approval plugin.
- MCP or npm plugin for Pi?
MCP (typesafe-mcp) exposes evaluate tools to any MCP host. Pi-specific packages like pi-jev-auto-mode and pi-warden hook Pi's extension API directly. Pick MCP when you want one server for multiple clients; pick Pi packages when you live entirely inside Pi.
- How fast should agent tooling feel?
Sub-second per gate is the design target for tool approvals. Published benchmarks in individual READMEs vary by model route and hardware; always measure on your network before you enable auto-approve in production.
- Do these tools replace my system prompt?
Rarely. They add a decision layer with explicit thresholds. Policy text still belongs in trusted channels (Hermes smart_policy, pi-warden.md rules, jev-shield JSON policies).
- Where is the official TypeSafe SDK?
Official JavaScript and Python clients live under Official and SDKs. Community MCP and Pi wrappers in this hub are maintained by their authors, not TypeSafe.
All listings
Agent toolingTypeScript2,735
Featured
Claude Code plugin that replaces the compaction summary with Jev decisions: every tool call and result is scored in one fast request, stale ones are dro
agent toolingcompactionagents
Agent toolingPython506
Demo
Software Factory Foreman: an agent supervisor that uses Jev decisions to keep coding agents on task.
agent-toolingagentssupervisor
Agent toolingTypeScript247
A staged code-review workflow and local dashboard built with TypeSafe Jev.
by @0xLogicrw
agent toolingagents
Agent toolingTypeScript141
TypeSafe Jev as a decision layer for the Pi coding agent: a measured tool-call gate plus jev_ask for typed, calibrated answers.
agent-toolingagentspi
Agent toolingJavaScript427
npm jev-router: jev-claude and jev-codex wrap real CLIs; Jev picks the cheapest sufficient model each fresh user turn with status or commentary lines.
agent-toolingroutingclaude-code
Agent toolingTypeScript113
Local-first MCP plugin for continuous software-quality review by AI coding agents, powered by Jev.
agent toolingmcpagents
Agent toolingMarkdown131
Agent skill teaching Choice, Score, Noul question design, state shape, and thresholds for jev-1.13 integrations.
agent-toolingskillagents
Agent toolingTypeScript316
MCP
Ten MCP tools on TypeSafe Jev: verify, screen, find, rerank, classify, decide, compare, extract, review, and gate in about 150 to 500 ms.
agent-toolingmcpagents
Agent toolingGo62
MCP
Go CLI and single-binary MCP server exposing TypeSafe judgments to Claude Desktop, Claude Code, and Codex.
by @0xLogicrw
agent toolingmcpagents
Agent toolingTypeScript61
Guardrails for Pi built on pi-typesafe that steer the agent instead of interrupting you: Jev judges irreversible and off-task tool calls, detects stuck
agent toolingguardrailsagents
Agent toolingTypeScript427
Demo
Search the web with TypeSafe's Jev: source selection, query understanding and relevance ranking.
agent-toolingragsearch
Agent toolingPython27
Per-turn model & reasoning routing for Codex, driven by Jev (TypeSafe System One): picks the model, thinking depth and speed mode for every turn.
by @0xLogicrw
agent toolingroutingagents
agent toolingagents
Agent toolingTypeScript21
Route HTTP requests by meaning. A semantic router for Hono powered by Jev.
agent toolingroutingagents
Agent toolingPython13
A calibrated context sieve for Claude Code: every tool result is judged by a System One model before it enters context.
by @0xLogicrw
agent toolingagentseval
Agent toolingTypeScript12
Agent-ergonomic CLI for TypeSafe's Jev: fast calibrated judgments (pick, rate, check, rank, triage, guard) from the shell.
agent toolingguardrailsagents
Agent toolingTypeScript10
Context-pruning proxy for Claude Code and Codex: Jev judges which history is still needed, measured not claimed.
agent toolingagents
Agent toolingTypeScript9
Jev (TypeSafe System One) backed auto mode for the Pi coding agent: semantically auto-approves bash, write, and edit tool calls and fails closed when a
agent toolingagents
Agent toolingPython7
MCP
MCP server for TypeSafe Jev: typed classify, score, check, match and screen for any agent, with confidence on every answer.
by @0xLogicrw
agent toolingmcpagents
Agent toolingTypeScript7
Configurable semantic linting powered by Jev, with file-level NOUL judgments and a magic-strings plugin.
agent toolingagents
agent toolingagents
Agent toolingTypeScript6
Semantic tool routing and typed System One decisions for the Pi coding agent using TypeSafe Jev.
agent toolingroutingagents
Agent toolingJavaScript6
Automatic model routing for Pi using TypeSafe's Jev through Vercel AI Gateway.
agent toolingroutingagents
agent toolingagents
Agent toolingPython6
TypeSafe (Jev) skill routing for Hermes Agent: names the one skill worth loading, before the model call.
agent toolingroutingagents
Agent toolingPython5
Typed, confidence-aware agent skill routing with TypeSafe Jev.
agent toolingroutingagents
Agent toolingJavaScript5
MCP
Connect JEV to MCP clients and compare its judgments against general-purpose LLMs using shared datasets and measurable accuracy.
agent toolingmcpsdk
agent toolingroutingagents
Agent toolingTypeScript4
MCP
Brainwires jevwire packages a Jev decision layer for agents: Streamable MCP server, embeddable DecisionModel library, and a Claude Code plugin marketpla
agent toolingmcpagents
Agent toolingTypeScript3
Demo
Every code file in a pull request, judged against Uncle Bob's Clean Code by TypeSafe's Jev, then reviewed by Luna.
agent toolingagents
agent toolingroutingagents
Agent toolingPython3
Typed System One decisions, ranking, verification, and an opt-in Hermes tool gate using TypeSafe Jev.
agent toolingagents
Agent toolingPython3
PoC: TypeSafe Jev as the reviewer for Hermes Agent smart command approvals.
agent toolingagents
Agent toolingPython3
Minimal agent loop where Jev directs control flow and a LangChain chat model writes argument values and the final response.
agent toolingagents
agent toolingagents
Agent toolingPython3
A small, fast prose linter: ruff-style rule codes for writing, backed by TypeSafe's Jev model.
agent toolingagents
Agent toolingTypeScript3
Voice-driven semantic auto-advance for Slidev, powered by Cloudflare Agents and TypeSafe AI Jev.
agent toolingagents
Agent toolingPython3
CLI and agent skill for TypeSafe System One (Jev): typed Choice, Score, and Noul judgments.
agent toolingagents
Agent toolingTypeScript2
Hybrid coding harness: System 2 writes, System 1 (Jev) runs reflexes.
agent toolingagents
Agent toolingTypeScript2
Codex plugin: verbatim Jev-guided context restoration around session compaction. Port of tamaratran/fast-jev-compaction to Codex lifecycl...
agent toolingcompactionagents
Agent toolingTypeScript2
Local proxy that picks the Claude model and effort per message using TypeSafe Jev. Routes subagents, leaves your cached main chat alone.
agent toolingroutingagents
Agent toolingJavaScript33
Auto mode for every coding agent: Jev risk on each tool call, injection scan on results, skill checks; npm jev-guard with multi-host adapters.
agent-toolingguardrailsprompt-injection
agent toolingagents
Agent toolingTypeScript2
MCPDemo
MCP server exposing TypeSafe Jev as typed, calibrated judgment tools: classify, score, check, batched ask.
agent toolingmcpagents
Agent toolingJavaScript2
Semantic MCP firewall powered by Jev, screens every tool call, tool result, and tool description with calibrated System One verification.
agent toolingmcpguardrails
agent toolingagents
Agent tooling2
System-architecture skill for TypeSafe AI Jev/System One, find fuzzy semantic judgment and turn it into small Choice/Score/Noul primitives.
by @Kantorcodes
agent toolingagents
Agent toolingTypeScript2
Build versioned judgment functions on TypeSafe's Jev once, then call the same published version from your backend over HTTP and from coding agents over
agent toolingmcpagents
Agent toolingPython2
Grep, but the pattern is a description. Filters lines by meaning with TypeSafe's Jev decision model: ~200 ms and a thousandth of a cent p...
agent toolingagents
Agent toolingTypeScript2
Verbatim Jev-scored context reduction for omp, over TypeSafe or OpenRouter.
agent toolingroutingcompaction
Agent toolingTypeScript2
Pi extension: verbatim context compaction with TypeSafe Jev decisions.
agent toolingcompactionagents
Agent toolingTypeScript2
Runtime constraints for the pi coding agent: checks every side-effecting tool call against what you said, before it runs. Powered by Type...
agent toolingagents
Agent toolingPython2
Tell Claude Code which installed skill a session needs, using Jev (TypeSafe AI) for the decision and skills.sh for discovery.
agent toolingroutingagents
agent toolingroutingagents
Agent toolingPython2
⚡ Ultra-fast, low-cost intelligent task classifier and 3-tier routing engine powered by TypeSafe Jev (System One).
agent toolingroutingagents
Agent toolingTypeScript2
Automated database migration safety reviewer powered by TypeSafe AI (Jev System One model).
agent toolingguardrailsagents
Agent toolingTypeScript2
Claude Code mod that routes decisions to TypeSafe's Jev model: ranks installed skills per prompt, and answers the agent's own this-or-that questions whe
agent toolingroutingagents
Agent toolingPython1
Fast semantic code search & diff sanity auditor for AI coding assistants (Antigravity, Cursor, Claude Code) powered by TypeSafe System One.
agent toolingmcpagents
Agent toolingPowerShell1
Utilizing Jev, the RLCD-type model provided by TypeSafe AI, to independently and cheaply judge agentic coding sessions.
agent toolingagents
Agent toolingTypeScript1
Scans a codebase for covert, deceptive, or data-stealing behavior with Jev, then reports suspicious files and line ranges before the user...
agent toolingagents
agent toolingagents
Agent toolingTypeScript1
MCP
MCP server that puts TypeSafe Jev on the coding loop in Cursor, Codex, and any MCP client.
agent toolingmcpsdk
Agent toolingHTML1
Predict another skill's next closed decision with TypeSafe Jev, without running that skill.
agent toolingagents
Agent toolingPython1
Open-source LLM router that uses TypeSafe's Jev to pick a model, on top of LiteLLM.
agent toolingroutingagents
Agent toolingTypeScript1
Per-prompt capability router for coding agents: resolves installed skills, MCP servers, agents and commands against your prompt via TypeSafe Jev, and me
agent toolingroutingmcp
Agent toolingPython1
A Stop hook that stops your coding agent from stopping too early. Plain-language rules, judged by jev.
agent toolingagents
Agent toolingTypeScript1
TypeSafe AI (Jev) adversarial reviewer and typesafe_ask tool for the omp coding agent.
agent toolingagents
Agent toolingTypeScript
Reusable GitHub Action: agent fix loop gated by checks, an AI reviewer, and TypeSafe Jev.
agent toolingagents
Agent toolingPython
Experimental protocol for evidence-aware agent handoffs, with Jev-assisted review before results reach the lead agent.
by @zxdubx
agent toolingagents
Agent toolingTypeScript
A CLI and GitHub Action that assesses code-change risk using deterministic rules and TypeSafe Jev, recommending checks and reviewers befo...
agent toolingagents
agent toolingroutingagents
Agent toolingPython
Experimental Hermes plugin: Jev-assisted model routing plans with budget and capability constraints. API access pending.
agent toolingroutingagents
Agent toolingPython
Grok skill: Jev as a judgment sensor in a builder-agent loop (priors × probabilities → next act).
agent toolingagents
Agent toolingPython
Autonomous Jev pull-request review with typed decisions, calibrated approval gates, and trusted-owner escalation.
agent toolingmoderationagents
Agent toolingTypeScript
Uses typeful jev, zero sync to pull and sync large repositories for issue triage.
agent toolingagents
Agent toolingTypeScript
Jev-powered model-judged permission gate for OMP (TypeSafe System One).
agent toolingagents
Agent toolingTypeScript
Demo
Send Pi agents back to work when they stop before the job is done.
agent toolingagents
Agent toolingTypeScript
Single-agent Pi coding coprocessor with Jev semantic gates, baseline-to-current diff review, and append-only observability telemetry.
agent toolingagents
Agent toolingTypeScript
Pi coding-agent extension built on the TypeSafe AI System One API (Jev).
agent toolingagents
Agent toolingTypeScript
A 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 you
agent toolingagents
Agent toolingPython
Guardrail + model router for LLM gateways on TypeSafe's Jev (System One model), with an independent accuracy/calibration/latency evaluation.
agent toolingroutingguardrails
Agent toolingTypeScript
MCP server exposing TypeSafe System One judgments (noul, choice, score) as agent tools.
agent toolingmcpagents
agent toolingagents
Agent toolingTypeScript
Typed judgment layer for coding agents, gates from PRD to ship. Jev-ready, provider-agnostic.
agent toolingagents
agent toolingmcpclassification
Agent toolingTypeScript
Demo
jev-codes is a devtools project from Kushal Agarwal that points Jev at a pull request diff and a YAML standards pack.
agent-toolingreviewdevtools
agent-toolinghermesmcp
agent-toolingmemoryrag
Agent toolingTypeScript78
Supervisor hooks for Claude Code and Codex: append-only ledger, deterministic done-without-check blocks, optional Jev judgments that never sole-block.
agent-toolingguardrailscodex
Agent toolingTypeScript70
Demo
Pre-release CLI: policy and quota filters, then TypeSafe ranks agent, model, and effort; launches Cursor, Claude Code, Codex, or OpenCode via Herdr.
agent-toolingroutingherdr
Agent toolingRust688
Rust coding agent harness: Jev routes reasoning consults, effort, and utility summaries; fail-open; Distill owns tool permissions.
agent-toolingroutingrust
Agent toolingPython287
Sutro CLI: label uncertain rows, GEPA-optimize TypeSafe Jev AI Functions, ship portable definitions via ai-functions.dev.
agent-toolingcligepa
Agent toolingPython601
Nokia research: Decider API turns hub LLMs into calibrated Jev-style choices with L0 flip fixes and optional L2 heads on vLLM.
agent-toolingopen-sourcedecision-model
Agent toolingTypeScript771
GitHub Next Bun server: local POST /v1/systemone over oMLX DiffusionGemma chat JSON, TypeSafe SDK compatible, README honest on calibration.
agent-toolingself-hostedsystem-one
Agent toolingPython606
SGLang JSON Schema generation on existing AR LLMs: enums, numbers, depends_on graphs, JevBench 228/231 with thinking. Not a System One server.
agent-toolingsglangjson-schema
Agent toolingPython255
Demo
LLM writes task harness code, freezes it, then Jev makes fast choice/score/noul calls. GEPA evolution and Pokémon 25% to 75% Eval demo.
agent-toolingharnessgepa
Agent toolingRust153
Ollama for decision models: pull laya kev von decider, serve POST /v1/systemone on :11435, ONNX graphs only, TypeSafe SDK drop-in.
agent-toolingself-hostedonnx
Agent toolingTypeScript1,148
jevgrep is for coding agents that need repo search, not another line grep. The jg CLI from dzhng/jevgrep uses Jev to walk folders, files, and declarations, then prints verbatim excerpts. A README eval on 10 Python SWE-bench tasks kept 8 of 10 solves and measured lower total agent cost than search without it.
agent-toolingclicode-search
Agent toolingTypeScript670
LLM plus Jev alloy: JevSift, JevLoop, and JevLong over a CLI JSON protocol so the host agent keeps the session.
agent-toolingharnesscli
Agent toolingTypeScript100
Automatic Claude Code, Cursor, and Codex project memory: JEVMEM.md with Jev decide, recall, guard, and poisoning checks; npm jevmem 0.6.1.
agent-toolingmemoryclaude-code
Agent toolingTypeScript336
Coding agent with a judgment kernel: a small judge answers about 35 bounded calls per turn while the big model does the work. CLI and desktop, built on pi and AionUi.
agent-toolingcoding-agentpi
Agent toolingTypeScript147
Ten levels from one Noul smart-if to a pi agent that writes its own Jev questions, with a live Vue lab and a YouTube walkthrough.
agent-toolingcurriculumpi
Agent toolingJavaScript98
Claude Code skill that hands bulk filter, classify, rank, find, and ask calls to TypeSafe Jev, with a 12-task README bench claiming 86% fewer Claude tokens.
agent-toolingclaude-codeskill
Agent toolingGo55
Unix-style thinking coreutils in Go: filter, label, sort, and cut by meaning with TypeSafe Jev, and print the user's own lines back.
agent-toolingcliunix
Agent toolingGo125
Tree-sitter unit linter: plain-language rules judged by Jev per function, type, or statement, with packs, cache, and Release binaries.
agent-toolinglintcli
Agent toolingPython111
On-prem clinical note extraction: local noul, choice, and score variables on one GPU, built on Jared Palmer kev with Augmented Clinical Notes training data.
agent-toolinghealthcaregpu
Agent toolingPython95
Find the page that answers a question in a long PDF: PageIndex builds a tree, Jev Choice walks it, and Noul checks the pages.
agent-toolingdocumentssearch
Agent toolingTypeScript301
Local proxy for Codex, Claude Code, OpenCode, Kilo, Gemini, and Devin: Jev picks the tool each turn, fails open, listens on 127.0.0.1 only.
agent-toolingcodexclaude-code
Agent toolingTypeScript185
Demo
DeepSeek Harness plugin: twelve optional Jev or Luna judgment features for skill ranking, drift, completion, approvals, and log admission.
agent-toolingharnessguardrails
Agent toolingGo158
Go CLI: multi-provider web search for coding agents, Jev scores and dedupes hits, optional scrape chunk filter; skimmer splits the same filter for any file or URL.
agent-toolingsearchcli
Agent toolingPython64
PyPI MCP and CLI: Jev scores every link and passage on a site crawl so the agent LLM only reads verbatim blocks that answer the goal.
agent-toolingmcpcrawler
Agent toolingPython105
PyPI jevals: Ragas-style and agent evals in one Jev request per trace, plus YAML gates. Openlayer ships bench tables versus Ragas and LLM judges.
agent-toolingevalguardrails
Agent toolingPython31
MIT MCP and CLI: 1 fps contact sheets, one Jev request per minute to find video moments, verify on full frames, about $0.02 to $0.03 per hour screened.
agent-toolingmcpvideo
Agent toolingTypeScript89
Demo
Proactive pi-durable news desk: cron sweeps 28 sources, Jev gates relevance and Hype/Facts/Signal, Vue browser plus TUI and CLI over one API.
agent-toolingproactive-agentnews