CLI Reference
A complete guide to all available plb CLI commands, their usage, arguments, and options.
Project & Integrity
Commands for initializing a project and managing its integrity.
init
Initializes a new Prompt Lockbox project in the specified directory. This command creates the prompts/ and .plb/ directories, along with the plb.toml and .plb.lock configuration files.
The directory to initialize the project in. Defaults to the current directory.
status
Displays a table showing the lock status and integrity of all prompts in the project. The output shows whether prompts are Locked, Unlocked, TAMPERED, or MISSING.
lock
Validates and locks a prompt to ensure its integrity. This records the prompt’s secure hash in the .plb.lock file.
The name, ID, or file path of the prompt to lock.
unlock
Unlocks a previously locked prompt, allowing it to be edited. This removes its entry from the .plb.lock file.
The name, ID, or file path of the prompt to unlock.
verify
Verifies the integrity of all locked prompts against the lockfile. Ideal for use in CI/CD pipelines.
This command will exit with a non-zero status code if any locked prompt is found to be tampered with or missing, causing a CI/CD job to fail.
lint
Validates all prompts in the project for schema compliance, correct syntax, and consistency.
Prompt Management
Commands for creating, viewing, and managing individual prompts.
list
Lists all prompts in a table format.
Displays a wider table with more details, including intended model and default inputs.
show
Displays the full metadata and template content for a single prompt.
The name, ID, or file path of the prompt to display.
create
Interactively creates a new prompt file, or creates template files in bulk.
If specified, creates a number of blank prompt template files non-interactively.
When run without options, plb create starts an interactive wizard to guide you through creating a single, detailed prompt.
run
Renders a prompt with provided parameters and displays the final output.
The name, ID, or path of the prompt to run.
Options:
Pre-sets a template variable. Can be used multiple times.
Opens your system’s default text editor to fill in all required prompt inputs.
After rendering, sends the prompt to the configured AI provider and displays the response.
version
Creates a new, version-bumped copy of a prompt. Defaults to a minor bump.
tree
Displays a hierarchical tree of prompts based on their namespace.
Searching & Indexing
Commands for building search indexes and finding prompts.
index
Builds a search index for all prompts to enable advanced search methods.
The indexing method to use. Choices are hybrid or splade.
search
A command group for searching prompts using different methods.
Subcommands (METHOD)
fuzzy: Lightweight search. No index required.hybrid: Balanced keyword and semantic search.splade: State-of-the-art sparse vector search.
AI Features
AI-powered commands for automating prompt engineering tasks.
prompt document
Automatically generates a description and tags for one or more prompts using an AI.
One or more prompt names, IDs, or paths to document.
Options:
Document all prompts in the project. Overrides any specific identifiers.
prompt improve
Gets an AI-powered critique and suggested improvements for a prompt.
This concludes the CLI command reference for plb. For additional help, use plb <command> --help, or explore related guides in our documentation to deepen your understanding and streamline your workflow.