CLI¶
Reversa has a simple CLI to manage the installation and lifecycle of agents in your project. All commands run with npx reversa in the project root.
Initial behavior¶
When the CLI starts and before it shows the Reversa ASCII logo, it must clear the terminal screen. The logo should appear at the top of the terminal, with no previous content above it.
The by sandeco signature must appear in white on the last line of the artwork, after a right-side margin from the end of the large Reversa word. It must not float in the middle of the logo height.
Expected format:
______
| ___ \
| |_/ /_____ _____ _ __ ___ __ _
| // _ \ \ / / _ \ '__/ __|/ _` |
| |\ \ __/\ V / __/ | \__ \ (_| |
\_| \_\___| \_/ \___|_| |___/\__,_| by sandeco
AI-Powered Reverse Engineering Framework
Available commands¶
install¶
Installs Reversa in the current legacy project. Detects present engines, asks for your preferences, and creates the entire required structure.
Use once, in the root of the project you want to analyze.
Installation Menu Layout¶
The installer must treat the menu as the main interface, not as a text dump. Questions must be numbered, have a blank line before the question, and, when options are shown, a blank line between the question and the list.
After the user confirms a multi-select question, the CLI must not print every selected item in one continuous line. This is forbidden because it creates a long, unreadable paragraph. Use one of these alternatives:
- Do not render the full selection and continue to the next question.
- Render a short summary, one line per team.
The agents menu lists teams, not individual agents. The user picks at the team level; the installer expands each selected team into its agents:
Reversa Agents Core(rendered in gray as a separator, always installed)Migration AgentsCode Forward AgentsPricing and Size AgentsTranslators N8N->Specs->Python(unchecked by default)
Reversa Agents Core is rendered as a gray, non-selectable separator that visually shows (*) as if it were a checked-and-disabled item: the user sees it, knows it is included, and the cursor skips over it. It contains all discovery agents (Reversa, Scout, Archaeologist, Detective, Architect, Writer, Reviewer, Visor, Data Master, Design System, Agents Help, Reconstructor), so the previous "Discovery Add-ons" group no longer exists as a separate concept. Even though the menu hides the agent-level detail, the final installation summary still breaks the count down by team (Discovery, Migration, Code Forward, Translators, Pricing).
status¶
Shows the current analysis state: which phase is in progress, which agents have already run, what's left to complete.
Useful for a quick overview before resuming a session.
update¶
Updates agents to the latest version of Reversa.
The command is smart: it checks the SHA-256 manifest of each file and never overwrites files you've customized. If you made adjustments to any agent, they stay intact.
add-agent¶
Adds a specific agent to the project. Useful if you didn't install all agents during the initial installation and now want to include, for example, Data Master or Design System.
add-engine¶
Adds support for an AI engine that wasn't present when you installed. For example: you installed only for Claude Code and now want to add Codex.
uninstall¶
Removes Reversa from the project: deletes the files created by the installation (.reversa/, .agents/skills/reversa-*/, engine entry files).
Your files stay intact
uninstall removes only what Reversa created. No original project file is touched. Specifications generated in _reversa_sdd/ are also preserved by default.