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

@agentproto/adapter-knowledge-files

v0.2.2

Published

Pure, zero-runtime knowledge adapter: implements @agentproto/knowledge-engine's IKnowledgeProvider over local files with an in-process BM25 index (no embeddings, no network, no API key). Registers as a provider-kit family and ships a node:fs FsPort so it

Readme

@agentproto/adapter-knowledge-files

The pure, zero-runtime files backing for the @agentproto/knowledge-engine IKnowledgeProvider contract.

Knowledge retrieval from local workspace files via an in-process BM25 index — no embeddings, no vector store, no network, no API key. It works the instant a workspace has files to index, which makes it the natural default corpus backing where AIP-10 entries already live as markdown: the entries are the index, with no separate store to keep in sync.

What's in the box

| Export | Role | | --- | --- | | FilesKnowledgeAdapter | The IKnowledgeProvider implementation. Consumes an injected FsPort; knows nothing about node:fs or process.env. | | LocalFs | A node:fs-backed FsPort (from @agentproto/corpus) so the adapter runs standalone. Paths are workspace-relative and can't escape the root. | | createStandaloneFilesAdapter() | Wires LocalFs + the typed env into a ready-to-use adapter. | | BM25 (tokenize, buildIndex, buildIndexYielding, score) | The dependency-free ranking core. Confined to this package. | | provider-kit family (KNOWLEDGE_FILES_CATALOG, makeKnowledgeFilesResolver, resolveKnowledgeBackend) | Registers the files backend under the @agentproto/adapter-knowledge-* discovery convention. |

Usage

Standalone, driven by env:

import { createStandaloneFilesAdapter } from "@agentproto/adapter-knowledge-files"

const kb = createStandaloneFilesAdapter()
const { hits } = await kb.query({ query: "usage based pricing", topK: 5 })

With a host-provided FsPort (e.g. a studio guild workspace):

import { FilesKnowledgeAdapter } from "@agentproto/adapter-knowledge-files"

const kb = new FilesKnowledgeAdapter({ fs, workspacePath: "knowledge" })

Through the provider-kit resolver (discovery / catalogs):

import { makeKnowledgeFilesResolver } from "@agentproto/adapter-knowledge-files"

const handle = await makeKnowledgeFilesResolver()("files")
const provider = handle?.provider()

Environment

Read once through the typed loadFilesKnowledgeEnv() — the only place these names live:

| Var | Default | Meaning | | --- | --- | --- | | KNOWLEDGE_FILES_ROOT | process.cwd() | Absolute host path the LocalFs is rooted at. | | KNOWLEDGE_FILES_PATH | knowledge | Workspace-relative folder walked and indexed (recursively). |

Indexing model

  • Indexed extensions: .md, .markdown, .txt, .yaml, .yml, .json. YAML frontmatter is stripped before tokenizing so schema fences don't pollute the vocabulary; frontmatter is still surfaced as each hit's metadata (and a slug mirrors to entrySlug for corpus hydration).
  • Ranking: standard BM25 (k1=1.5, b=0.75) with +1-smoothed IDF.
  • Cache: the index is built lazily on first query and cached permanently — it's rebuilt only when the adapter invalidate()s (via ingest/delete) or a fresh instance is created. Hosts wanting second-precision freshness on external edits can wire fs-watch into a new instance.

Capabilities

vectorSearch: false, hybridSearch: false, citations: true — BM25 returns ranked hits with a file path + byte span, usable as citations even without a vector concept.

Apache-2.0.