IntelliShell Documentation

For variables that require real-time data (like branch names or container IDs), you can define Dynamic Completions. A completion is a shell command that runs in the background to generate a list of suggestions on the fly.

Common use cases include:

• Listing Git branches for git checkout {{branch}}.
• Listing Docker containers for docker exec -it {{container}} bash.
• Fetching Kubernetes services for kubectl logs {{service}}.

In auto mode, you can exclude results containing a specific word by prefixing it with !.

Example: docker !test will find commands matching "docker" but excluding any that contain "test".

You can customize the system prompts for the suggest, fix, import, and completion tasks in the [ai.prompts] section. Use these dynamic placeholders to provide context to the AI:

• ##OS_SHELL_INFO##: Details about the current operating system and shell.
• ##WORKING_DIR##: The current working directory path and a tree-like view of its contents.
• ##SHELL_HISTORY##: The last few commands from the shell's history (available only for the fix task).

Any word in a command's description starting with # is treated as a searchable tag.

To filter by hashtag:

1. Press <kbd>Ctrl</kbd>+<kbd>Space</kbd> to open the search UI.
2. Type # to see a list of available hashtags.
3. Selecting a hashtag will filter the command list to only show templates containing that tag.

Tip: Hashtag discovery is cumulative. Searching for docker #compose and then typing # again will only suggest tags that appear on commands matching both docker and #compose.

IntelliShell allows you to transform static commands into interactive builders using two main mechanisms:

1. Variables: Wrap parts of a command in double curly braces {{variable_name}}. When you recall this template, IntelliShell will prompt you to provide values for these variables.
2. Dynamic Completions: You can power these templates by defining custom scripts. These scripts run to generate live, interactive suggestions for your variables (for example, a script that returns a list of active Docker containers to populate a variable).

To prevent sensitive information like API tokens, passwords, or comments from being saved to your suggestion history, wrap the variable in triple curly braces: {{{variable_name}}}.

Behavior:

• The value is used in the command exactly once.
• The value is never saved to suggestion history.
• If a matching environment variable exists (e.g., PASSWORD), IntelliShell will suggest using the environment variable (e.g., $PASSWORD or $env:PASSWORD) instead of the plain text value to prevent exposure.

curl -u user:{{{password}}} https://api.example.com/data

IntelliShell shifts the command-line paradigm from manual execution to a structured 'library of intents'. Instead of typing raw commands, you build and use reusable, interactive 'recipes' (command templates). This allows you to:

• Organize workflows: Store commands in a searchable knowledge base with descriptions and tags.
• Automate complex steps: Create dynamic commands that fetch and inject data (like IDs) automatically.
• Reduce friction: Use short searches or aliases instead of re-typing long, complex strings.
• Prevent errors: Minimize typos by reusing tested templates.

IntelliShell facilitates faster onboarding by providing instant command discovery and executable documentation:

• Instant Discovery: By including a .intellishell file in the repository, new developers can press ctrl+space to immediately view all essential project commands without searching through README files.
• Executable Documentation: Instead of static text, IntelliShell provides ready-to-run commands that include descriptions and placeholders for arguments, acting as living documentation for project setup, builds, and deployments.

IntelliShell identifies and visually highlights potentially dangerous commands (e.g., using red warning styles) to prevent accidental execution. This is handled through two methods:

1. Tag-Based Detection: Explicitly tagging a command with #destructive will always flag it.
2. Config-Based Regex Detection: You can define custom regular expression patterns in your configuration file under the [destructive].patterns section to catch other dangerous commands.