Eru.Tool 0.13.0

There is a newer version of this package available.
See the version list below for details.
dotnet tool install --global Eru.Tool --version 0.13.0
                    
This package contains a .NET tool you can call from the shell/command line.
dotnet new tool-manifest
                    
if you are setting up this repo
dotnet tool install --local Eru.Tool --version 0.13.0
                    
This package contains a .NET tool you can call from the shell/command line.
#tool dotnet:?package=Eru.Tool&version=0.13.0
                    
nuke :add-package Eru.Tool --version 0.13.0
                    

eru

eru is a CLI tool for sharing knowledge files between projects. Declare where your shared files live (a git repo), pull them in with a single command, and track everything in a lock file so they stay in sync.

  • Manifests — knowledge-source repos publish .eru/manifest.json to declare available files; consumers pull from it automatically
  • Collections — group related files into a named set and pull them all with one command
  • Tag-based pulls — --tag devops fetches everything tagged devops across all collections
  • Glob patterns — collection entries can use globs to pull multiple files in one reference (e.g. docs/*.md)
  • Lock file (.eru/eru.lock) records every pulled file: its origin, path, and content hash
  • Static site — eru site generate builds a self-contained HTML site for browsing and searching the local knowledge cache; eru site serve adds live reload and a search API
  • Dry-run mode on every write command
  • Global config (~/.config/eru/config.json) for sources and collections shared across all your repos

Requirements

Install

Install the Eru.Tool tool:

dotnet tool install --global Eru.Tool

Leave off --global to install locally in a repo instead.

Prebuilt self-contained binaries (linux-x64, win-x64, osx-x64) are also attached to each GitHub release if you'd rather not install the .NET SDK.

Install the skill with skills.sh:

npx skills add dburriss/eru

The skill tells agents how to use the CLI.

Quick start

1. Initialise your repo

eru init

Creates .eru/config.json in the current directory.

2. Pull a file

Paste a GitHub or GitLab file URL directly:

eru add https://github.com/my-org/knowledge/blob/main/docs/adr-template.md

eru registers the source automatically, downloads the file, and records it in .eru/eru.lock.

Or pull by source name and path once a source is configured:

eru add knowledge:docs/adr-template.md

Or pull an entire curated collection:

eru add --collection onboarding-docs

3. Keep files up to date

eru sync

Fetches every file in .eru/eru.lock, compares content hashes, and overwrites anything that has drifted.


Commands

Command What it does
eru init Scaffold .eru/config.json in the current directory
eru init --global Create the global config at ~/.config/eru/config.json
eru add <path> Pull a file by source:path, bare filename, or full URL
eru add --collection <name> Pull all files in a named collection
eru add --tag <tag> Pull all files matching a tag
eru add --dryrun <path> Preview what would be pulled
eru search <terms> Search sources and the lock file
eru sync Re-fetch all tracked files and update drifted ones
eru sync --dryrun Preview what sync would change
eru source add <url> Register a git repo as a knowledge source
eru source list List configured knowledge sources
eru source view <name> Show details and files for a source
eru source files <name> List the files a source exposes
eru source bundle add Register a knowledge bundle (manifest or OKF) on a source
eru source remove <name> Remove a source
eru collection create <name> Create a new collection
eru collection add <name> -f <source:path> Add a file reference to a collection
eru print <source:path\|hash> Print a document's contents to stdout
eru config list\|view\|get\|set\|migrate Inspect, edit and migrate configuration
eru remove <path> Delete a pulled file and drop it from the lock file
eru disconnect <path> Stop tracking a file but keep it on disk
eru browse Browse the knowledge cache in an interactive terminal UI
eru cache prune / eru cache clear Manage the local knowledge cache
eru inbox add/list/remove/default Manage inboxes (local, or remote git repos)
eru inbox channel add/list/remove Manage channels and their ACP agents
eru inbox send Send an item to an inbox
eru inbox process / eru inbox watch Run inbox items through the channel's agent, once or continuously
eru manifest init Create .eru/manifest.json in a knowledge-source repo
eru manifest add <path> Add a file/glob entry to the manifest
eru manifest remove <path> Remove an entry from the manifest
eru manifest validate Check all manifest entries resolve to local files (alias: verify)
eru site generate [-o <dir>] Generate a static HTML site for browsing the local knowledge cache
eru site serve [-p <port>] Serve the site locally with live reload and a search API
eru okf validate <path> Check a directory tree for OKF §11 conformance, and warn on broken links (--strict-links to fail)
eru okf init <path> Create missing OKF index.md files
eru okf fix <path> Repair a directory tree so it passes OKF validation
eru okf verify <file> Record a human/machine verification on a concept file
eru version Print the version and git commit
eru mcp Start an MCP stdio server for AI agent use

For full argument details see docs/reference/cli.md.

Static site

eru site generate builds a self-contained HTML site from the local knowledge cache — browse and search all source files offline.

eru site serve does the same but also starts a local HTTP server with a live-reload SSE feed and a /api/search endpoint. The browser reconnects and reloads automatically whenever the cache is synced. See docs/reference/site-generation.md and docs/how-to/customize-the-generated-site.md.

MCP server

eru mcp exposes knowledge search and retrieval to AI agents (Claude, Copilot, Cursor, etc.) over the Model Context Protocol. See docs/how-to/set-up-the-mcp-server.md.

Contributing

Run mise install for the toolchain, then dotnet build and dotnet test --solution eru.slnx. CI builds and tests every push and PR; pushing a v* tag publishes the NuGet package and GitHub release binaries. See AGENTS.md for the project layout, conventions and release flow.

Documentation

Full documentation lives under docs/, organised by Diataxis: tutorials, how-to guides, reference, and explanation.

Product Compatible and additional computed target framework versions.
.NET net10.0 is compatible.  net10.0-android was computed.  net10.0-browser was computed.  net10.0-ios was computed.  net10.0-maccatalyst was computed.  net10.0-macos was computed.  net10.0-tvos was computed.  net10.0-windows was computed. 
Compatible target framework(s)
Included target framework(s) (in package)
Learn more about Target Frameworks and .NET Standard.

This package has no dependencies.

Version Downloads Last Updated
0.16.1 93 11/7/2026
0.16.0 80 11/7/2026
0.15.1 89 11/5/2026
0.15.0 83 11/5/2026
0.14.0 116 11/2/2026
0.13.1 143 10/12/2026
0.13.0 124 10/12/2026
0.12.0 132 10/9/2026
0.11.1 125 10/8/2026
0.11.0 127 10/8/2026
0.10.0 131 10/5/2026
0.9.1 146 10/2/2026
0.9.0 138 10/2/2026
0.7.1 350 5/29/2026
0.7.0 273 5/29/2026
0.6.0 269 5/29/2026
0.5.1 266 5/29/2026
0.5.0 262 5/29/2026
0.4.0 264 5/29/2026
0.3.0 264 5/28/2026
Loading failed

`eru config` inspects and manages configuration: `list` shows the global and local config files with their paths and versions, `view` shows the effective config with the origin (`local`, `global` or `default`) of each value, `get <key>` shows one setting, `set <key> <json-value> [--global] [--unset]` edits a setting (validated before writing, files left untouched on error) and `migrate [--local|--global] [--dryrun]` rewrites an older config as version 3, keeping a `.bak` of the original
Configurable generated site. Every page is built from four slots (`top`, `left`, `right`, `main`) of widgets, set under `site.layout` (site-wide) and per page; a slot is an ordered list or `false`. Widgets: `breadcrumbs`, `title`, `meta`, `verified`, `generated`, `description`, `tags`, `stale`, `cli`, `content`, `list`, `toc`, `related`, `graph`, `facets`, `links`
`site.pages` defines custom pages at `pages/<id>/index.html`: a `filter` (by `source`, `bundle`, `type`, `tag`, `status`, `ext`, `path` globs, `stale`, with `not`) with `sort`, `group`, `limit` and `open` (`page`, `resource` or `source`), a `content` page rendering a repo Markdown file (`content.file`) or an indexed document (`content.doc`), or both. The built-in pages (`browse`, `sources`, `tags`, `types`, `bundles`) can be dropped with `false` or overridden by a page of the same id, and `site.home` picks the page at `index.html`
Three list views, `cards`, `list` and `table`, chosen with `site.defaultView` or a page's `view`; `site.cards.columns` (1-6) sets the card grid's maximum columns. Table headers sort when JavaScript is available
`site.document` configures document pages: `layout` for every document and `rules` that match on `type`, `tag`, `bundle`, `source` and the other filter fields, so different documents get different layouts. The `verified` widget shows the trust tier with each verifier and date; `generated` shows who or what generated a document and when
`eru site serve` rebuilds when a local Markdown file used by a content page changes
Site config errors (unknown page or widget, a widget in a slot it is not valid in, a bad view or column, a missing or out-of-repo content file) fail with a message naming the key; a failed rebuild writes nothing, so `eru site serve` keeps serving the last good build
**Breaking:** the generated site has refreshed defaults. Every page gets a facet column on the left and, when it has headings or links, a table of contents on the right; lists are a card grid of up to three columns; the top nav is built from the pages that exist. Document pages are rebuilt from widgets, and the default document layout shows the link graph at the top of the page and the linked-document lists at the bottom. To get close to the old look, set `site.layout.left` and `site.layout.right` to `false` and `site.cards.columns` to `1`
Browse filtering is driven by markup instead of per-list ids: any element marked `data-doc` is filterable, `data-<facet>` attributes hold its values as JSON arrays and `data-facet="<name>"` on a sidebar list makes it a filter. Types and bundles are now checkbox facets on Browse. Custom CSS or scripts that relied on `#file-list`, `#file-count` or the `source-filters`, `ext-filters` and `tag-filters` ids need updating
The page-level `SiteFeatures` flags `TagPages` and `SourcePages` are removed from `GenerateOptions`; drop pages with `site.pages.<id>: false` instead (`types` and `bundles` were previously tied to `TagPages`)
**Breaking:** config files are now version 3. Settings are grouped into top-level sections, shared by `~/.config/eru/config.json` and `.eru/config.json`, and the `settings` / `defaults` blocks are removed. Section names are singular and the entities a section holds always live under `items`. Version 1 and 2 files are migrated automatically when read and rewritten as version 3 only on the next save (read-only commands leave the file untouched); newly created configs are written as version 3
Every settings section is now valid in both the global and the local file, merging local-over-global per key. `sync.stateFile` was previously local-only, and `sync.branch` and `mcp.refreshIntervalMinutes` global-only
`eru inbox default` writes `inbox.default`
Reading a version 1 config (sources with `basePath`) migrates `basePath` to a `manifest` bundle instead of dropping it