Get started

Choose a Lock

One Lock per kind of data source. Add each one to a project with lox add <lock>, then keep lox run open.

Your agent needs…	Lock	What it does
A SQL database (PostgreSQL, MySQL, …)	locksql	The agent queries, PII comes back masked; connections and writes wait for you.
RabbitMQ	lockmq	Masked queue inspection; connections, peeks and unmasked reads wait for you.
Files on S3, SFTP or disk	lockfile	The agent asks yes/no questions about files it never reads.
The web or an HTTP API	lockhttp	Reads of trusted hosts pass; anything that sends data or reaches an unknown host waits for you. Credentials are injected by the proxy, never seen by the agent.
Logs	locklog	Masked, bounded log search; the agent reads the logs, not the people in them.
Commands on a server	lockssh	Allow-listed commands run at once, the rest waits for you; the SSH key never leaves the console.

Note

lockhttp and lockssh are pre-releases, locklog an early release. lockssh is not in the Homebrew tap yet. The overview page of each Lock says what its current release supports.

Every Lock works the same way #

Credentials stay in the Lock. Its console holds the password, key or token, in its own account and container.
Masked by default. The agent sees masked values; the hub only ever sees masked screens.
You sign outside the harness. lox run approves by rule or asks you.
Every decision is audited. lox audit lists each exchange.

Each Lock documents what the agent can see and its residual risks: see Security.