~/wiki/ideas
Tools and integrations: MCP servers, skills, clients, gateways and libraries
TL;DR Official tools first; the rest is unaudited community code, mostly days old (★ and push as of 2026-09-23). Vet per Community repos: what people built and how they use Jev before granting shell, browser or key access.
Pattern IDs: Decision patterns from the community (with fit verdicts). Licence MIT unless stated. Numbers: the author's.
| Your job | Look at |
|---|---|
| Guard a coding agent | skills, plugins, hooks |
| Give an agent Jev as a tool | MCP servers |
| Email, helpdesk, CRM, ops | n8n node, Composio |
| E-commerce store | Magento 2 module |
| Rank, filter or lint from code or SQL | CLIs, libraries |
| Call Jev from Go, R, PHP, Rust | SDKs table |
| Run a Jev-like model locally | open replicas (not TypeSafe) |
Not captured: Shopify, Gorgias, Zendesk, Gmail, WooCommerce integrations.
Official (TypeSafe)
- Python
typesafe-sdk→ Python SDK: install, clients, system_one(); JS@typesafe-ai/sdk→ JavaScript/TypeScript SDK quick contract: install, client, choice/score/noul, answers, errors (the page a builder reads); skill and Claude Code plugintypesafe-ai/skills→ The typesafe-ai agent skill and Claude Code plugin. system-one-adapter, an LLM-backed stand-in → system-one-adapter: LLM-backed drop-in for TypeSafeClient. All repos: typesafe-ai GitHub organisation and repos.
MCP servers
| Tool | Exposes | Install | ★ · push | Note |
|---|---|---|---|---|
| jev-scout (kierandotai) | guarded search, fetch, report; scores relevance, credibility, steering; SSRF guard | clone, claude mcp add |
0 · 2026-09-21 · no licence | 88% relevance, 96% credibility on a labelled set. P17 |
| BrowserClaw | Chrome extension, 49 tools; a goal loop decides with Jev, escalates destructive keywords | release zip | 6 · 2026-09-23 · NOASSERTION | your logged-in Chrome. P12 |
| Jevbridge (tacticocc) | jev_decide, jev_gate, jev_computer_use + ACP; LLM or heuristic stand-in without a key |
clone | 40 · 2026-09-21 | stand-in lacks Jev's calibration. P03, P14 |
| tenbin (Shingo Imota) | design-time MCP + skill: decompose, lint, measure thresholds, rerank, cost guard | clone | 2 · 2026-09-21 | production calls the SDK |
MCP also in: jev-rs, jev-browser, codex-context-diet, JevGuard. Elsewhere: jev-mcp, typesafe-mcp, skillbox → Repos: coding agents, orchestration, memory, browser and computer use, integrations.
Agent skills, plugins and hooks
| Tool | Does | Install | ★ · push | Note |
|---|---|---|---|---|
| Bouncer (Clownware) | Claude Code PreToolUse: 7 Nouls, YAML policy, 15 hard rules first; observe default |
plugin marketplace | 1 · 2026-09-20 · no licence | 104 near-miss fixtures, Brier per question; cat .env scored 0.15, so rules own the edges. P03 |
| jev-gates | 7 Claude Code gates (rule, scope, intent, done, claims, proof, commit); escalate, never approve | plugin | 0 · 2026-09-20 | P04, P10 |
| jev-belay | Stop hook; asks only if files changed and no check passed | plugin | 18 · 2026-09-20 | AUROC 0.976 on 100 stops; 346 ms; $0.00005. P04 |
| Abide (coldteadotai) | one question per AGENTS.md rule per edit; Claude Code, Codex, OpenCode | npx @coldtea/abide init |
210 · 2026-09-20 | 93 sessions: 39 edits flagged, 10 confirmed. P10 |
| codex-context-diet | Codex PostToolUse: 5 questions per bulky result; stale ones cut to head + note; local Laya option |
plugin | 3 · 2026-09-23 · NOASSERTION | 20 of 65 results cut; ~$0.00003 each. P07 |
| Yoshi (CompozyOS) | proxy for Claude Code and Codex; Jev picks history to omit | Bun | 25 · 2026-09-18 | both current trials slower, no validated saving. P07 |
| Jevvy (PanAchy) | auto-approves harmless shell asks, else abstains | npx @jevvy/permissions init |
15 · 2026-09-21 | 5 providers incl. custom endpoint. P03 |
| Interlock (somoore) | gate(): detectors → Noul battery → policy code; agent sees only canary secrets |
pip install -e . |
0 · 2026-09-19 | 38-case regression set, not a benchmark. P03 |
| Switchloom (Instructa) | Codex skill routing tasks across models, Jev mode | cargo install |
1 · 2026-09-20 | author advises against: no Jev gain in two pilots. P02 |
| Hunch (Kelbie) | plain-English repo search, diff review rules, GitHub Action | npx skills add Kelbie/hunch |
4 · 2026-09-23 | TypeSafe, gateway or local SemIf. P10, P17 |
| Jevify (altryne) | offload bulk judgments; "Jevify this" audits a product | npx skills add altryne/jevify --skill jevify |
25 · 2026-09-21 | not ryana/jevify |
| Jev Architect (Karan Bansal) | find, design, test decision loops; may say "keep it in code" | npx skills add or plugin |
6 · 2026-09-20 | cf. Consult guide: could Jev help this project? |
| prompt2jev (sumleo) | prompt or LLM code → linted request.json + script |
plugin or npx skills add |
3 · 2026-09-21 | P01 |
| clay-jev-people-ranker (PromptGTM) | Clay filters, Choice + Noul qualify founders | install_skill.py |
0 · 2026-09-21 | 577 people in 20.1 s; 490,395 tok ≈ $0.0206 (= $0.042/Mtok). P19 |
Elsewhere: building-with-jev-skill, ryana/jevify, hermes-jev-skills, fast-jev-compaction, winnow, Canny, pi-warden, jev-codex-router, jev-router, compact-adviser, jev-rules → Repos: coding agents, orchestration, memory, browser and computer use, integrations.
SDKs, clients and framework integrations
| Tool | Does | Install | ★ · push | Note |
|---|---|---|---|---|
| typesafe-sdk-go (Radix ILS) | Go client on Python SDK 0.6.0's wire format | go get |
0 · 2026-09-18 | TYPESAFE_* names verified |
| laravel-typesafe-jev | Laravel facade, async, Jev::fake(); unofficial PHP SDK |
composer require |
3 · 2026-09-17 | |
| jevr | R client jev_ask(); TypeSafe or OpenRouter |
pak::pak() |
4 · 2026-09-20 · NOASSERTION | |
| Vercel AI SDK (guide) | experimental_evaluate, ai ≥ 7.0.105, typesafe-ai/jev |
npm i ai |
Vercel doc | limits, price verified; its boolean = API noul; per-request ZDR unverified |
| Effect, Composio | @effect/ai-typesafe; @composio/typesafe, composio-typesafe |
npm / PyPI | 16184, 30293 · 2026-09-23 | READMEs only list them |
| n8n node (n3ndor) | multi-question node, usable as an agent tool | Community Nodes | 2 · 2026-09-19 | P22 |
| home-assistant-typesafe (Allen Porter) | HA conversation agent, gate 0.40, fallback agent | custom component | 0 · 2026-09-21 · Apache-2.0 | P14 |
| Magento 2 module (@hckhenrique) | one call per order, customer, product, review, cart | composer (VCS) | 1 · 2026-09-21 · NOASSERTION | P18 |
| MinusPodJev (@ttlequals0) | OpenAI-compatible proxy, one Noul per podcast segment | Docker | 1 · 2026-09-22 | POC. P27 |
| tripwire (noelzappy) | AI SDK middleware or proxy, 7 checks per LLM reply | npm i @noelzappy/tripwire |
2 · 2026-09-18 | no accuracy vs real Jev yet. P16 |
Elsewhere: HA-Jev, pg-jev → Repos: data, documents, judging, real-time, markets, business apps, replicas; langchain-typesafe → Measurements, access routes and open replicas.
Gateways and access routes
- OpenRouter (
~typesafe/jev-latest), Vercel AI Gateway (typesafe-ai/jev):verifiedin SDK 0.7.1 viabase_url/TYPESAFE_BASE_URL(Python SDK internals: headers, key validation, gateways and base URLs, logging, forward compatibility, dependencies, migration notes); their prices and terms apply. - New API plugin (@t0ng7u) and OpenCode Zen (Jevvy; jev-agent-skill, model
jev-1.13-freeat $0):unverified. - impossibl serves Laya, not Jev, on a compatible endpoint.
CLIs, libraries and dev tools
| Tool | Does | Install | ★ · push | Note |
|---|---|---|---|---|
| jev CLI (model-clis) | one-question commands, --assert → exit 3, presets, map over JSONL |
brew, scoop, curl | 0 · 2026-09-18 | exit 0 ≠ "yes"; ~20 req/s = 1,200 rpm verified. P10 |
| jev-axi (Nicholas Underwood) | pick/rate/check/rank/diff/triage/guard; setup safety hook |
npm i -g jev-axi |
19 · 2026-09-19 | file ranking did not cut agent cost. P03, P11, P16 |
| jevkit (Ariel Frischer) | Rust CLI; jev lint: 13 offline rules |
install.sh | 3 · 2026-09-19 | |
| jevcache (hyperspaceai) | cache on redacted (model, schema, state); CI replay |
curl | 71 · 2026-09-19 · no licence | answers not bit-exact (Measurements, access routes and open replicas) |
| JevGuard (Seb4Ez) | cache, UNRESOLVED_OR_OTHER option, AMBIGUOUS_STATE flag |
pip install jevguard-core |
0 · 2026-09-22 | its "official" MCP is not TypeSafe's |
| jevcal (abhixhek) | thresholds per accuracy target; CI drift | pip from git | 10 · 2026-09-18 | MCA benchmark-ban claim contradicts docs since 2026-09-19 (Legal: MCA, DPA, privacy, data retention). P06 |
| jev-tree (reachjalil) | recursive Choice past the 255 cap (verified) |
npm install jev-tree |
6 · 2026-09-18 | Cookbook: Hierarchical classification |
| jev-reranker (shinpr) | JSON pipe: rerank, filter, extract | npm i -g jev-reranker |
4 · 2026-09-23 | P17 |
| duckdb-jev (Colliber) | answers as ENUM, DOUBLE, STRUCT |
DuckDB extension | 23 · 2026-09-18 | P17, P18 |
| pg_typesafe (Giulio Piccolo) | Postgres functions, batched *_many |
make install |
82 · 2026-09-20 | 38 texts: 23 s vs 0.86 s batched. P17 |
| jev-lint (mizchi) | ast-grep + one question; 99 rules | npx -y jev-lint check |
74 · 2026-09-22 | P10 |
| oxlint-plugin-jev (wobsoriano) | Oxlint rules as English yes/no | npm i -D oxlint oxlint-plugin-jev |
58 · 2026-09-21 | P10 |
| slop-grader (lukstei) | every rule × every line, feeds an agent fix plan | npx @lukstei/slop-grader |
20 · 2026-09-22 | P30 |
| doc-router (misbahsy) | which PDF pages need OCR | cargo | 26 · 2026-09-18 | 155 → 87 pages billed; 9 misses vs 28 for rules. P26 |
| fastbrowse (Agent Labs) | Jev picks actions, LLM reads, cited answers, --authorize gate |
uvx fastbrowse |
98 · 2026-09-22 | 41/42 vs Browser Use 39/42; median $0.0041 vs $0.3668/task. P12 |
| jev-browser (Tomohiko Hiraki) | Playwright SDK, CLI, MCP | release tarball | 7 · 2026-09-22 · Apache-2.0 | P12 |
| System One Harness | finite actions → questions, gated steps, traces | pip install -e . |
124 · 2026-09-21 · Apache-2.0 | P24 |
| evoke | sentence → program + args; runs, confirms or asks | curl; npm | 11 · 2026-09-23 · Apache-2.0 | P14 |
| firehose-judge (Leo Mata) | Cloudflare template: 8 questions per post, human lane | wrangler deploy |
0 · 2026-09-19 | ~260 ms, ~850 tokens/post. P16, P06 |
Elsewhere: semdecide, blink, jev-shell-history → Repos: coding agents, orchestration, memory, browser and computer use, integrations.
Open replicas and Jev-compatible servers
Trust note. None is TypeSafe or runs Jev. /v1/systemone compatibility is the wire shape, not the calibration Confidence-gated routing needs; several compute their own confidence. Re-measure on your labels (Measurements, access routes and open replicas).
| Tool | What | Install | ★ · push | Note |
|---|---|---|---|---|
| Laya (Nandakishor M) | 3 encoder checkpoints + router; laya-serve |
pip install laya |
18593 · 2026-09-23 · Apache-2.0 | 32.8 ms on a T4; Jev figures quoted, not measured; Jev leads Banking77 0.870 vs 0.425; Khmer 0.000 at 0.952 confidence |
| Von (wfzyx) | 395M encoder | Hugging Face | 526 · 2026-09-23 · Apache-2.0 | own table: Jev 96.6% vs Von 72.0% on 49 tasks |
| AnyJev (Nokia Applied Research) | any HF LLM, debiased, optional head on 100–300 labels | pip install "anyjev[hf]" |
267 · 2026-09-23 · Apache-2.0 | order flips 0.230 → 0.073 (Qwen3-8B) |
| Simple Jev (Eugene Cheah) | logits server, free demo API | pip from repo | 491 · 2026-09-21 · Apache-2.0 | own contract |
| jev-rs (Yijun Yu) | Rust over llama-server; serve, MCP, eval | cargo install jev-rs |
5 · 2026-09-22 · Apache-2.0 | SDKs reach it via TYPESAFE_BASE_URL |
| hearim (ziozzang) | Go gateway over Ollama, vLLM, SGLang | go build |
5 · 2026-09-22 · NOASSERTION | entropy-based confidence |
| openJev-verdict-2.0 (Heman10x) | 149.6M ModernBERT, confidence head, WebGPU | Hugging Face | 273 · 2026-09-20 · NOASSERTION | claims 77.10% vs Jev 72.70%, Jev's figure copied from Laya, not measured |
| visual-jev | image decisions, Qwen3-VL | uv | 2 · 2026-09-20 · no licence | Jev is text-only (Models, aliases, pricing, rate limits, context) |
| JevTuner | Brier-loss training recipe | clone | 12 · 2026-09-21 · no licence | not TypeSafe's method |
| Recipe (@theanandprasad) | softmax over yes/no logits | — | post 2026-09-22 | 21/22 on Llama 3.3 70B; raw logits flip with option order (AnyJev) |
Elsewhere: SemIf, NanoJev, jevlike, openjev-sglang, mini-jev → Repos: data, documents, judging, real-time, markets, business apps, replicas.
Related
- Community repos: what people built and how they use Jev — vetting; Repos: coding agents, orchestration, memory, browser and computer use, integrations, Repos: data, documents, judging, real-time, markets, business apps, replicas — repos in depth
- Measurements, access routes and open replicas — gateway and replica trust; TYPESAFE_* environment variables across SDKs — base URL and keys
Sources
Links inline; raw captures (2026-09-23) in frontmatter.