AI Summaries
CommandTree uses GitHub Copilot to automatically generate a one-sentence, plain-language summary for every discovered command. When GitHub Copilot is installed, hover over any command in the tree to see exactly what it does — and get warnings about dangerous operations.
How It Works
After CommandTree discovers your commands, it sends each script's content to GitHub Copilot and asks for a one-to-two sentence description. These summaries appear in the tooltip when you hover over a command.
Summaries are stored in a local SQLite database at .commandtree/commandtree.sqlite3 in your workspace root. They persist across sessions and only regenerate when the underlying script changes (detected via content hashing).
Security Warnings
Copilot also analyses each command for potentially dangerous operations like rm -rf, git push --force, or credential handling. When a risk is detected, the command's label is prefixed with a warning indicator and the tooltip includes a security warning section.
Requirements
- GitHub Copilot extension installed and signed in
- The
commandtree.enableAiSummariessetting enabled (on by default)
If Copilot is not available, CommandTree works exactly as before — all core features (discovery, running, tagging, filtering) are fully independent of AI summaries.
Triggering Summaries
Summaries generate automatically on activation and when files change. To manually regenerate, run the CommandTree: Generate AI Summaries command from the command palette.
Frequently Asked Questions
What does an AI summary look like?
Each summary is a one-to-two sentence plain-language description of what the command does. For example, a shell script that runs database migrations might show: "Runs pending database migrations and seeds the development database." Hover over any command in the tree to see its summary.
Are summaries stored locally?
Yes. All summaries are stored in a SQLite database at .commandtree/commandtree.sqlite3 in your workspace root. No data is sent to external servers beyond the GitHub Copilot API that runs locally in VS Code.
How are security warnings triggered?
Copilot analyses each command for potentially dangerous operations such as rm -rf, git push --force, file permission changes, or credential handling. When a risk is detected, the command label shows a warning indicator and the tooltip explains the specific risk.
Can I disable AI summaries?
Yes. Set commandtree.enableAiSummaries to false in your VS Code settings. All other features — discovery, execution, tagging, and filtering — work independently of AI summaries.