OpenClaw gets a full Russian-language guide to its JSON5 config, CLI and secure setup

A detailed guide to OpenClaw has appeared on Habr. OpenClaw is a platform where almost all operational setup hangs on a single file, openclaw.json, and a large CLI. The article gathers in one place what users previously had to dig out of the official help, GitHub issues and community posts.

What's in the config

The author explains how the top level of the configuration is organized and why OpenClaw demands care when you edit it by hand. The file uses JSON5, so it allows comments and trailing commas, but the schema is validated very strictly at startup. If you add an unknown key, the gateway simply won't start. The guide separately describes gateway, env, agents, models, auth, channels, session, cron, tools, sandbox and logging, as well as how hot reload works without a full restart.

• gateway: port, bind mode, access token and hot reload behavior
• agents and models: base settings, fallbacks and provider selection
• auth and channels: profile switching, Telegram, Discord and other channels
• session and cron: conversation isolation, time-based resets and background jobs
• env and secrets: variable precedence, SecretRef and storage of sensitive data

The main value of the piece is that it goes beyond a list of fields. The text explains how the different configuration levels relate, where defaults apply, when settings take effect without a restart and which parameters are better left untouched by hand. For anyone running OpenClaw not locally but on a VPS or in a production environment, this is no longer a reference but a full operational playbook.

Where setup breaks

The most useful section deals not with the happy path but with typical failures. The guide stresses that some parameters work only in agents.defaults and are silently ignored if moved into the settings of a specific agent. As a result, a config can look correct while the desired behavior never turns on. The recommendation is simple: use the openclaw config get, set and unset commands more often instead of editing directly, and run diagnostics after every change.

openclaw doctor is the most useful command in the entire CLI.

The guide also covers the problem of context bloat. OpenClaw uses two different mechanisms: pruning removes old tool results before a model call, while compaction compresses history when the context window is nearly full. As an example, it cites a case where 35 messages turned into a 2.9 MB session file and a context of 208,467 tokens, causing the system to stop responding with a model limit of 200 thousand tokens.

CLI for admins

The CLI is described by working scenario rather than alphabetically, which makes the guide practical. Commands for starting the gateway, daemon status, channel management, logs, cron jobs, plugins, skills, sessions and updates are gathered in one place. For infrastructure commands, it separately shows safe modes: health works even without a running daemon, security audit checks the config for risks, and secrets audit helps catch credential leaks before production.

The last important section covers environment variables and version control. The guide fixes the order of precedence: first the process environment, then the local .env, then the global .env in the OpenClaw directory, and only after that the built-in env.vars. For secrets, it advises using references to variables or SecretRef rather than values in openclaw.json, and committing only a configuration template to the repository, excluding .env, session folders, memory files and temporary state.

What this means

OpenClaw is becoming a tool not only for enthusiasts willing to read the source code, but also for teams that need a clear operational playbook. If a project relies on several models, channels and automation, a guide like this lowers the risk of silent breakages, leaked secrets and unstable behavior after updates.