Tag: claude-code

An agent skill is a folder with a SKILL.md file that the agent loads on demand: only its name and description sit in the system prompt, the instructions arrive when a task matches, and bundled scripts run without ever entering context. Here is how that differs from a system prompt in cost, scope, invocation and trust, with the exact limits from the Agent Skills spec and Claude Code.

MCP is a JSON-RPC protocol that lets any AI host call tools, read resources and use prompts from any server, the way LSP did for language support. Here is how it works on the wire under revision 2026-07-28, captured from a real server, and why VS Code, Visual Studio, JetBrains, Xcode, Zed and Cursor all converged on it.

A coding agent refactors a component, renames a data-testid, and your Playwright suite goes red. Here is a three-layer setup, verified on Playwright 1.63.0 and Claude Code 2.1.197: locators that survive, a PreToolUse hook that blocks the rename, and a CI guard that lets the Playwright healer agent fix locators but never rewrite an assertion.

Every agent request pays for its system prompt, the provider's hidden tool-use preamble, and every tool schema before the user says anything. How to measure that fixed cost with differential count_tokens calls on the Anthropic API, the input_tokens endpoint on OpenAI, an MCP listTools probe, and /context in Claude Code, anchored to claude-opus-5-5, claude-sonnet-5, and @modelcontextprotocol/client 2.1.0.

Dev containers, Docker Sandboxes microVMs, Lima VMs, and hosted cloud sandboxes all move Claude Code, Codex, or Copilot CLI off your host. What each one actually isolates, the exact commands to make it throwaway, and the two things none of them protect: your writable workspace and your git hooks.

Scoping a token does not stop a prompt-injected agent from leaking it. Moving the token out of the agent's process does. Three working setups: a 50-line gateway any agent can call, Claude Code 2.1.199+ sandbox masking with injectHosts, and Claude Managed Agents vaults, plus the measured leak that host-level injection still allows.

Claude Code 2.1.277 reads AGENTS.md natively, so Cursor, Claude Code, and most Copilot surfaces can now share one instructions file. Here is the checklist for merging three drifted rule files into AGENTS.md, which bridge files to delete, the one Copilot file you still have to generate, and a CI check that keeps them from drifting again.

A side chat branches off what the agent already knows; a subagent goes and finds out something new. Claude Code's /btw has no tools at all, Cursor's /btw opens a full agent, and a plain subagent sees none of your conversation. A decision guide for Claude Code 2.1.278 and Cursor 3.11+.

Move a local stdio MCP server to a remote Streamable HTTP server behind OAuth on MCP spec 2026-07-28 and TypeScript SDK 2.0.0. Dynamic Client Registration is now deprecated, so use Client ID Metadata Documents first and keep DCR only as a fallback. Covers the steps, a measured CIMD/DCR test matrix, and the misleading error you get when the two sides disagree.

A September 2026 comparison of GitHub Copilot code review, Cursor Bugbot, and anthropics/claude-code-action@v1 running the code-review plugin: what each one reads, what it skips, what it costs per review, and which to put on which repo.