Tag
13 posts tagged.
You gave your agent a shell tool, used execFile so there is no shell, and allowlisted the programs. It is still not safe. If the agent controls the argument array, tar's own --checkpoint-action=exec runs any command with no metacharacters. Here is the reproduction and the per-binary argument policy that stops it, measured on Node 24.8.0.
If your agent has a read_file tool, you probably guard it with path.resolve plus startsWith. It blocks the obvious traversal and is defeated by a symlink and by a sibling directory sharing the prefix. Both reproduced on Node 24.8.0, with the guard that passes all nine cases.
Set _meta["anthropic/requiresUserInteraction"] to true and Claude Code prompts for that tool on every call, even in bypassPermissions, and even when an allow rule matches. It shipped in v2.1.199 with no changelog entry, which is why almost nobody has written about it.
A runnable Python guide to Claude structured output: the native output_config.format JSON path, Pydantic-driven schemas, strict tool use, a portable tool-use fallback, and the 400/truncation gotchas nobody covers.
The five parts every AI agent has, a runnable Python example in about 40 lines, and an honest guide to the guardrails and framework choices most tutorials skip.
A runnable Python quickstart: install the Claude Agent SDK, stream a run with query(), then give the agent your own tools with the @tool decorator and ClaudeSDKClient. Builds a weather agent that chains two tools.
A runnable 2026 quickstart: install the Claude Agent SDK, wire a custom tool with tool() and createSdkMcpServer(), and let the agent loop call it for you in about 40 lines of TypeScript.
Should you use the Claude Agent SDK or hand-roll the agent loop with the Anthropic Client SDK? An honest 2026 comparison with runnable TypeScript for both and a decision table.
Speed up your Claude agent loop by running independent tool calls concurrently in TypeScript. Copy-paste code for parallel tool_use, is_error handling, and disable_parallel_tool_use.
A complete TypeScript tutorial for the streaming agent loop on Claude: input_json_delta accumulation, multi-turn dispatch, AbortController cancellation, and the eager_input_streaming workaround for the verified 5 second first-content delay on tool use. About $0.03 per call with claude-sonnet-4-6 at June 2026 pricing.
An AI agent is just a loop: you call a model, the model asks to run a tool, you run it, you feed the result back, and you repeat until the model is done. In this tutorial you build that loop yourself in plain TypeScript against the Anthropic Messages API, no framework. You will wire up two tools (read a file, run a calculation), let the model orchestrate them, add a turn cap and basic guardrails, then verify the whole thing end to end. The result is a small research agent you fully understand and can extend with your own tools.
You do not need to rebuild your product to ship an AI agent inside it. The trick is to expose the service functions you already have, search records, create an order, fetch a customer, as tools, then run a small server-side agent loop that the model uses to orchestrate them. This tutorial wraps an existing service layer as tools, scopes every call to the authenticated user, separates safe read tools from gated write tools, exposes the agent as one authenticated endpoint, and deploys that endpoint to Totalum. Your database, auth, and business logic stay untouched.
