Claude Code 2.1.292 lets you choose effort for each subagent
Claude Code 2.1.292 adds effort to the Agent tool. Use it to give routine searches less work and difficult reviews more, then check the results.
Blog
Plain-English posts on running coding agents in isolated boxes: safety, parallel sessions, snapshots and forks, verification on a real desktop, and how Sessionboxer compares with Devin, Cursor, Codex, Claude Code and OpenHands.
How to keep your secrets away from a coding agent that runs without prompts
An agent with its prompts off can read any file on its disk and send it anywhere. How to keep .env files, keys and logins out of its reach.
Devin vs Cursor Cloud Agents vs Codex cloud vs Claude Code on the web vs OpenHands, compared (October 2026)
Six ways to run a coding agent in its own machine: where it runs, whose account pays, whether you can watch the screen, snapshot the box or switch agents.
How to run Claude Code with --dangerously-skip-permissions, safely
Claude Code's docs say to skip permission prompts only inside a container or VM. What the flag turns off, what still protects you, and how to set up that box.
A coding agent deleted a production database in 9 seconds. What a sandbox stops, and what it does not
In April 2026 a Cursor agent wiped a startup's database and backups during a staging task. The lesson is not more prompts. It is what the agent can reach.
Claude Code mods are not sandboxed. Where to run them
Claude Code 2.1.287 added mods: JavaScript hooks that rewrite prompts and tool calls and run with your permissions. That changes where the agent should run.
Which AI coding agent should you use? A guide by who you are
Solo developer, tech lead, QA engineer, on-call SRE, student or manager: the right coding agent setup is different for each. A plain guide to picking yours.
How to run several coding agents in parallel without them stepping on each other
Worktrees, containers or cloud VMs? How to run two, five or ten coding agent sessions at once, what breaks, and how to keep the reviews manageable.
Let the coding agent test its own work on a real desktop (and send you the video)
Devin 2.2 and Cursor Cloud Agents test their changes with computer use. Get the same loop on your own machine, after every turn, with a captioned recording.
Snapshot and fork a coding agent's whole machine, then hand it to a different agent
Git saves your files. A snapshot saves the agent's machine: files, packages, browser state and its memory. Fork it, or hand it to a different agent.
An open-source, self-hosted alternative to Devin: what you actually need
The parts a self-hosted agent workspace needs to replace Devin, how OpenHands and Sessionboxer cover them, and where the gaps are.
What your coding agent really sends to the model: context, tokens and cost, made visible
Each Claude Code turn sends far more than your message. See the system prompt, tool schemas, MCP servers and history that fill the context window, and the cost.