by dragthelake
A menu bar app that keeps a written record of what you worked on.
# Add to your Claude Code skills
git clone https://github.com/dragthelake/ambient-contextGuides for using ai agents skills like ambient-context.
ambient-context is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by dragthelake. A menu bar app that keeps a written record of what you worked on. It has 161 GitHub stars.
ambient-context's catalog security scan is still queued. You can run an instant dependency and prompt-injection check now with the "Scan for vulnerabilities" button above.
Clone the repository with "git clone https://github.com/dragthelake/ambient-context" and add it to your Claude Code skills directory (see the Installation section above).
ambient-context is primarily written in Rust. It is open-source under dragthelake on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh ambient-context against similar tools.
No comments yet. Be the first to share your thoughts!
Unlocks once the catalog security scan passes (runs nightly).
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.
The deep catalog scan for this skill is still queued. Run an instant dependency check now instead.
Ambient Context writes down what you work on during the day to give your AI assistant better context.
A macOS menu bar app that keeps a written record of what you work on, in plain markdown, in a folder you own. Point Claude Code, Cursor, Codex or any MCP client at your context folder.
You spend the day reading, writing, replying and deciding. None of it reaches the AI assistant you open at 4pm. So you paste in the email, explain the ticket, summarise the call, and the model still works from a fraction of what you actually saw. The context existed. It was on your screen. Nobody wrote it down.
While the eye in your menu bar is open, the app reads the text of whichever window you have focused (through the macOS accessibility tree, every few seconds) and appends it to markdown files for the day. At a time you choose, or when you press the button, an agent you already have turns that record into two more things: a small cited knowledge base and a written summary of the day.
Three files per day, each one useful on its own:
| What it is | Who it is for | |
|---|---|---|
| Context | The record. What was on screen, when, in which app, with the file path or URL where the app exposes one. | Agents that need evidence. Grep. You, when you cannot remember which tab it was in. |
| Knowledge | Six cited pages built from the record: People, Commitments, Threads, Products, Issues, Reading. Every claim points back to a block in the record. | Agents building memory about your projects and the people in them. |
| Notes | The day written up from the knowledge base, with citations checked against the record before it is saved. | You, at the end of the day. Your standup. The model you open tomorrow morning. |
With Claude Code you can send the long input (the day's record) to a cheaper Context model and reserve a stronger Notes model for the page you will actually read.
Raw screen text is noisy. Ambient Context is built so an LLM can read it cheaply.
messages.md so the timeline stays readable and message bodies are easy
to exclude from a prompt.AGENTS.md in the folder explains the format to whatever reads it.The record is yours and it stays on your computer.
[redacted]. These are heuristics, not a guarantee, so treat the folder
as sensitive.The trust boundary, data inventory, claims matrix and known gaps are in Privacy and security.
Requires macOS 14 or later on Apple Silicon.
Download the signed, notarised DMG from
Releases,
drag Ambient Context.app to Applications and open it.
To build it yourself you need Node, Rust and the Xcode Command Line Tools:
git clone https://github.com/dragthelake/ambient-context
cd ambient-context
npm install
npm run tauri build
The app lands in src-tauri/target/release/bundle/macos/. For development,
npm run tauri dev runs it with hot reload.
The app binary is also an MCP server. For Claude Code:
claude mcp add --scope user --transport stdio ambient-context -- "/Applications/Ambient Context.app/Contents/MacOS/ambient-context" mcp
Claude Desktop, Cursor, Zed and generic stdio clients are covered in docs/mcp.md. Read tools work whether or not the app is running. Anything that changes state goes through the running app, so every write lands in the day's ledger with your client's name on it.
Days/2026-09-04/
apps.md the timeline, with native app bodies
websites.md a table of visits, without page bodies
messages.md mail and chat bodies, routed here
KB/2026-09-04/
people.md commitments.md threads.md products.md issues.md reading.md
manifest.md
Summaries/2026-09-04.md
AGENTS.md
cd src-tauri && cargo test
cd .. && npx tsc --noEmit && npx vitest run
Ambient Context is free, open source and built by one person. If it earns a place in your menu bar, three things help: star the repo, report the apps that come back thin, and if you want to, buy me a coffee.
MIT. The bundled Funnel fonts carry their SIL OFL text. See CHANGELOG.md for what is in each rele