Skip to main content

​
Overview

Enterprise Search lets you query all your company’s tools simultaneously. One question, all sources, synthesized results — no more tab switching between Slack, Gmail, Google Drive, and Notion. Primarily designed for Cowork, Anthropic’s agentic desktop application, but also works in Claude Code.

​
How It Works

Claude decomposes your question, runs targeted searches across every connected source in parallel, and synthesizes results into a single coherent answer with source attribution. 
You: "What did we decide about the API redesign?"
              ↓ Claude searches in parallel
~~chat: #engineering thread from Tuesday with the decision
~~email: Follow-up email from Sarah with the spec
~~cloud storage: Updated API design doc (modified yesterday)
              ↓ Claude synthesizes
"The team decided on Tuesday to go with REST over GraphQL.
 Sarah sent the updated spec Thursday. The design doc
 reflects the final approach."
No tab switching. No remembering which tool has what. Ask the question, get the answer.

​
Installation

Claude plugins add knowledge-work-plugins/enterprise-search

​
What It Searches

Connect any combination of sources. The more you connect, the more complete your answers.

Communication

Chat (~~chat) — Messages, threads, channels, DMs
Email (~~email) — Emails, attachments, conversations

Documents

Cloud Storage (~~cloud storage) — Docs, sheets, slides, PDFs
Wiki / Knowledge Base — Internal docs, runbooks

Work Tracking

Project Management — Tasks, issues, epics, milestones
Ticketing — Support tickets, customer issues

Business Data

CRM — Accounts, contacts, opportunities
Office Suite — Documents, spreadsheets, presentations
Each source is an MCP connection. Plugin files use ~~category placeholders (e.g., ~~chat, ~~email) that resolve to whichever MCP server you connect. See CONNECTORS.md for supported tools.

​
Commands

​
/search

Search across all connected sources in one query. 
/enterprise-search:search what's the status of Project Aurora?
/enterprise-search:search from:sarah about:budget after:2025-01-01
/enterprise-search:search decisions made in #product this week
Supported filters:
  • from: — Filter by sender/author
  • in: — Filter by channel, folder, or location
  • after: / before: — Date range filters
  • type: — Filter by content type (message, email, doc, thread, file)
Filters are automatically translated to each source’s native query syntax.

​
/digest

Generate a daily or weekly digest of activity across all sources. 
/enterprise-search:digest --daily      # What happened today
/enterprise-search:digest --weekly     # Weekly rollup grouped by project/topic
What it includes:
  • Action items — Direct requests, tasks assigned, questions for you
  • Decisions — Conclusions reached, approvals, policy changes
  • Mentions — Times you were mentioned or referenced
  • Updates — Status changes, completed items, document updates
Activity is grouped by topic/project, not by source, so you see what happened across all your tools in context.

​
Skills

Three specialized skills power the search experience:
Query decomposition and source-specific translation.What it does:
  • Analyzes your natural language question to understand intent
  • Breaks the query into targeted sub-queries for each source
  • Translates filters (from:, in:, date ranges) to each source’s native syntax
  • Handles ambiguity with clarifying questions
  • Falls back gracefully when sources are unavailable
Example:
Query: "What did Sarah say about the budget?"

Decomposed to:
- ~~chat: messages from:sarah containing "budget"
- ~~email: emails from:sarah subject:budget OR body:budget
- ~~cloud storage: docs authored_by:sarah containing "budget"
Manages which MCP sources are available and how to use them.What it does:
  • Detects which MCP servers are connected
  • Guides you to connect new sources when needed
  • Manages source priority for ranking results
  • Handles rate limits and failed connections
  • Prevents one failed source from blocking the entire search
Graceful degradation: If some sources fail, the search continues with available sources and notes what couldn’t be reached.
Combines results from multiple sources into coherent answers.What it does:
  • Deduplicates information appearing across sources (same decision in chat AND email = one entry)
  • Ranks results by relevance, freshness, authority, and completeness
  • Synthesizes raw results into direct answers, not just lists
  • Attributes sources so you can dig deeper
  • Scores confidence based on result quality
  • Summarizes large result sets
Ranking factors:
  • Relevance — How well does it match the query?
  • Freshness — Recent results rank higher for status/decision queries
  • Authority — Official docs > wiki > chat for facts; conversations > docs for “what did we discuss”
  • Completeness — Results with more context rank higher

​
Example Workflows

​
Finding a Decision

You: /enterprise-search:search when did we decide to switch to Postgres?

Claude searches:
  ~~chat → #engineering, #infrastructure for "postgres" "switch" "decision"
  ~~email → threads with "postgres" in subject
  ~~cloud storage → docs mentioning database migration

Result: "The decision was made March 3 in #infrastructure (link).
         Sarah's email on March 4 confirmed the timeline.
         The migration plan doc was updated March 5."

​
Catching Up After Time Off

You: /enterprise-search:digest --weekly

Claude scans:
  ~~chat → channels you're in, DMs, mentions
  ~~email → inbox activity
  ~~cloud storage → docs shared with you or modified

Result: Grouped summary by project with action items
        flagged and decisions highlighted.
        
# Weekly Digest — Feb 22-28, 2026

Sources scanned: ~~chat, ~~email, ~~cloud storage

## Action Items (3 items)
- [ ] Review budget proposal for Sarah — from Sarah, ~~email (Feb 27)
- [ ] Approve API spec v3 — from Greg, ~~chat (Feb 26)
- [ ] Q2 roadmap input needed by Friday — from Todd, ~~email (Feb 25)

## Decisions Made
- Switched to Postgres — #infrastructure (Feb 24)
- API v3 design approved — ~~email thread (Feb 26)

## Project Aurora
- ~~chat: Design review concluded, chose Option B (#design, Feb 23)
- ~~email: Sarah sent updated spec (Feb 24)
- ~~cloud storage: "Aurora API Spec v3" updated (Feb 24)
- 3 tasks moved to In Progress

​
Finding an Expert

You: /enterprise-search:search who knows about our Kubernetes setup?

Claude searches:
  ~~chat → messages about Kubernetes, k8s, clusters
  ~~cloud storage → docs authored about infrastructure
  Wiki → runbooks and architecture docs

Result: "Based on message history and doc authorship,
         Alex and Priya are your go-to people for k8s.
         Here's the main runbook (link)."

​
Complex Query with Filters

You: /enterprise-search:search from:sarah about:budget after:2025-01-01 in:#finance

Claude translates filters per source:
  ~~chat: from:sarah in:#finance after:2025-01-01 text:budget
  ~~email: from:sarah after:2025-01-01 subject:budget OR body:budget
  ~~cloud storage: author:sarah modified_after:2025-01-01 content:budget

Result: Unified results across all sources with timeline and context.

​
Connector Reference

CategoryPlaceholderIncludedAlternatives
Chat~~chatSlackMicrosoft Teams, Discord
Email~~emailMicrosoft 365—
Cloud storage~~cloud storageMicrosoft 365Dropbox
Knowledge base~~knowledge baseNotion, GuruConfluence, Slite
Project tracker~~project trackerAtlassian (Jira/Confluence), AsanaLinear, monday.com
CRM~~CRM(not pre-configured)Salesforce, HubSpot
Office suite~~office suiteMicrosoft 365Google Workspace
See CONNECTORS.md for details.

​
Getting Started

  1. Install the plugin 
    claude plugins add knowledge-work-plugins/enterprise-search
  2. Connect your tools via MCP
    The more sources you connect, the more complete your search results. Start with chat, email, and cloud storage.
  3. Search across everything 
    /enterprise-search:search [your question here]
  4. Get a digest 
    /enterprise-search:digest --daily

​
Philosophy

Knowledge workers spend hours every week hunting for information scattered across tools. The answer exists somewhere — in a Slack thread, an email chain, a doc, a wiki page — but finding it means searching each tool individually, cross-referencing results, and hoping you checked the right place. Enterprise Search treats all your tools as one searchable knowledge base. One query, all sources, synthesized results. Your company’s knowledge shouldn’t be locked in silos. Search everything at once.

Build docs developers (and LLMs) love

Get started for freeTalk to us