npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@trooth/semanto-verify

v1.6.0

Published

Check a Semanto signed copy, its response signatures and its transparency log proofs. No dependencies.

Readme

semanto-verify (JavaScript)

One dependency-free ES module, verify.mjs, for Node 20+ and browsers (WebCrypto only), plus cli.mjs.

node verifier/js/cli.mjs file verifier/test-vectors/v1/semanto.json --pages verifier/test-vectors/v1/pages
node verifier/js/cli.mjs page https://acme.com/security
node verifier/js/cli.mjs response https://acme.com/security
node verifier/js/cli.mjs log --from 1200 --to 1340

file needs no network when --keys is given or a keys.json sits beside the file (the vectors ship one). page fetches the URL with Accept: text/markdown, the JWK Sets from the domain and from Trooth, the _trooth DNS TXT record, and the proof the response links to with rel="proof", then prints every check in plain words. Exit code 0 means every check passed; 1 means at least one failed; 2 means the command was wrong.

In a browser:

<script type="module">
  import { verifyResponse } from "https://trooth.co/verifier/verify.mjs";
  const r = await fetch("/security", { headers: { Accept: "text/markdown" } });
  const body = new Uint8Array(await r.arrayBuffer());
  const keys = await (await fetch("/.well-known/trooth-keys.json")).json();
  const result = await verifyResponse(r.status, Object.fromEntries(r.headers), body, keys);
  console.log(result.ok, result.checks);
</script>

Functions: verifySemanto(file, pageBytesByUrl, keys, { now, revocations }), verifyPageProof(proof, pageBytes, keys, { now }), verifyResponse(status, headers, body, keys, { now }), verifyWebBotAuthRequest, verifyTreeHead, verifyInclusionInLog, verifyConsistency, verifyCheckpoint, verifyOffer, verifyRevocationList, stateOf(version, now, revocations), plus canonicalize (RFC 8785), sha256, the RFC 6962 primitives, verifyJws and verifyMessage (RFC 9421). Every result is { ok, reason, checks: [{ name, ok, detail }] } with the reason codes listed in verifier/test-vectors/v1/README.md. Type declarations are in verify.d.mts.

SemantoClient

client.mjs reads a page and checks it in one call: the markdown, both key sets, the response signature and the domain's _trooth-keys TXT record over DNS over HTTPS, with key sets and DNS answers kept five minutes.

import { SemantoClient } from "@trooth/semanto-verify/client";
const r = await new SemantoClient().read("https://acme.com/pricing");
if (r.ok) console.log(r.text); else console.log(r.reason, r.checks);

r.dns is true when the domain's DNS names the signing key, false when it names another, null when DNS could not be asked.

SemantoSource

source.mjs answers value questions from a company's own signed copy, in process. It loads /.well-known/trooth.semanto and the company's keys from the company's site, checks the file once for each new body, reads the site again at most every 15 minutes, and answers with the same matcher as Trooth's service (answer.mjs, whose answers are compared case by case with verifier/test-vectors/answer-v1).

import { SemantoSource } from "@trooth/semanto-verify/source";
const acme = new SemantoSource("acme.com", { onEvent: (e) => console.log(e.event, e.detail) });
const a = await acme.answer({ question: "How much is the Starter plan?" });
// a.status: "answer" | "candidates" | "not_published" | "not_read"; a.text is the signed sentence and the check link

For a chatbot, reply builds the text the customer sees by code and discards the model's draft: the signed sentence exactly as signed, its conditions, the day it was signed and the check link.

const r = await acme.reply({ question: userQuestion, draft: modelText });
show(r.text); // r.statement is the signed sentence; r.discarded says a draft was dropped

respond follows the company's answer mode, chosen on trooth.co: Checked builds the reply by code (as reply does), Source only sends the model's draft unchecked, and Audit sends the draft and checks it right after (r.audit resolves to supported, differs with the signed sentence, marked, not in record or not checked). Until the mode can be read, Checked is used.

const r = await acme.respond({ question: userQuestion, draft: modelText });
show(r.text);

Trooth's API is used only for Trooth's witness keys, the monitor's marks and the answer mode. When it cannot be reached, answers keep coming from the last version that checked out and the source logs trooth_api_unreachable. A new version that does not check out is refused (version_refused) and the last good one stays in use. A fact marked changed since signing or expired returns its mark and no value.

What the customer sees, turn by turn

reply returns what a company's own chatbot shows for one turn, built by code (turn.mjs):

| status | lines | | --- | --- | | answer | the signed sentence, its conditions, "From Acme Inc's signed record. Signed 2026-11-12. Check it: " | | not_published | the company's not-published line and its contact route from the record | | marked | the mark (changed since signing, or expired) and the contact route; no value | | choose | several signed answers, as up to three choices | | confirm | a locked value found only by word overlap: up to three choices to pick before any value | | closed | no checked version, a withdrawn version or a revoked key: no value, the contact route; onAlert is called once |

const r = await acme.reply({ question, history }); // history: earlier questions, used only to complete a follow-up
if (r.status === "confirm") showChoices(r.choices); // then acme.reply({ fact_id, confirmed: true })
else show(r.text);

outputStep({ question, text }) checks an agent's final answer and replaces it with the checked reply when it carries a value the company did not sign. quote({ question, select }) answers a free-form question with the company's own signed sentences, picked by ID by the company's model and printed by code. stream(tokens) releases a sentence only after its check. maskPersonal(text, file) masks emails, phone numbers and names that are not in the record.

Frameworks

frameworks.mjs adds the three modes and the output step to the Vercel AI SDK (semantoMiddleware), LangChain.js and LlamaIndex.TS (withSemanto), the OpenAI Agents SDK (semantoOutputGuardrail) and plain HTTPS (semantoHandler, a fetch handler). Sample chatbots are in verifier/examples/answer-path.