Local-first
Core workflows run on your machine. Any configured network service or publishing step is an explicit boundary.
Ecosystem
These three CLIs compose with GNO over JSON, skills, and MCP: Flow (Claude Code plans), sheets-cli (Sheets I/O), outlookctl (Outlook COM). Each is a separate MIT/open repo that speaks JSON and a skill or MCP interface.
Flow is a Claude Code plugin: research agents, gap analysis, then execution against a written plan.
Flow's research agents can query GNO via the MCP server or the agent skill. The context-scout and repo-scout agents pair with `gno query` for local hybrid retrieval while Flow keeps the plan.
/plugin marketplace add https://github.com/gmickel/gmickel-claude-marketplace/plugin install flowgno mcp install --target claude-desktopOmarchy shell plugin for GNO: a quiet bar widget with index health, an anchored panel with counts and recent documents, and a keyboard-first overlay (Super+R) for search, deep search, and opening documents.
Recall reads the same local index every other surface uses. Summon the overlay, search or deep-search the corpus, and open the document without leaving the desktop.
omarchy plugin add https://github.com/gmickel/omarchy-gno-recall --enableKey-based updates, batch ops, JSON in and out. Installs as a skill for Claude Code and Codex.
Export GNO search results to a tracking sheet. AI agents can query your knowledge base with GNO, then write structured findings to Google Sheets for downstream reporting or review.
gno query 'Q4 budget projections' --json | sheets-cli append --sheet 'Research Log'gno similar doc.md --json | sheets-cli upsert --sheet 'Related' --key idLocal CLI bridge for Outlook Classic automation via COM. AI-assisted email and calendar management from Claude Code. No API keys, no OAuth — just your existing authenticated session.
Find the right context in your GNO index before drafting a reply. AI agents can query your knowledge base to surface the meeting note, decision doc, or policy before composing an email.
gno query 'authentication decision' --answeroutlookctl draft --to team@company.com --subject 'Auth approach'The pattern
They compose because they share local-first defaults, JSON I/O, CLI+skill surfaces, and one-job tools.
Core workflows run on your machine. Any configured network service or publishing step is an explicit boundary.
Structured input and output for automation. AI agents parse results reliably, humans pipe them with jq.
Both human-usable at the terminal and AI-usable as a skill or MCP server. Same tool, two interfaces.
Pipe gno context build JSON into sheets-cli; draft from gno ask --verify with outlookctl.
Example workflow
# 1. Compile one bounded evidence Capsule
gno context build "performance optimization" --collection notes --budget 12000 --json --output capsule.json# 2. Write the evidence set to the tracking sheet
sheets-cli append --sheet Research --data @capsule.json# 3. Generate only what the evidence supports
gno ask "what did we learn" --verify --show-sources# 4. Draft the follow-up email
outlookctl draft --to lead@team.com --subject "Research"Build your own
Open an issue on the GNO repository. We love hearing about local-first, AI-native tools that compose with GNO’s CLI, REST API, or MCP server.