Term
Permission Mode
A global setting in Claude Code that controls how tool calls are handled: Plan (planning only), Default (asks for confirmation), Accept Edits (files free), Bypass (allow everything).
Permission Mode — explained in more detail
The permission mode is a global setting in Claude Code that applies to every tool call in a session. It defines when Claude may act without asking and when the user has to confirm. Four modes are available:
The four modes
- Plan: Claude only plans, no actions are executed. Ideal for reviewing an approach before anything changes.
- Default: For tools not on the explicit allowlist, Claude asks the user. The safest active mode for normal work.
- Accept Edits: File operations run without prompting, while riskier actions such as bash still require confirmation.
- Bypass: Everything is allowed, no prompts. Dangerous — only sensible in isolated sandboxes or throwaway environments.
Recommendation
Default for most everyday work. Accept Edits for longer, well-scoped sessions with many file changes. Bypass only in fully isolated environments where misbehaviour cannot cause real damage.
Entdecke mehr
Saving Tokens with Claude: 6 Principles That Make Experts Twice as Fast
How I turned my CLAUDE.md from a style guide into a token budget — 6 principles for lower cost, less waiting, and more honest reporting.
GlossarAuto-Classifier
A safety layer in Claude Code: a small LLM evaluates every tool call that is not explicitly allowlisted, live, and blocks destructive or unauthorised actions before they execute.
LexikonLangGraph
LangGraph as the standard for agent orchestration — nodes, edges, state, loops, human-in-the-loop and persistence explained clearly.