A shell command hook for Claude Code that auto-approves safe commands and blocks destructive ones to reduce permission fatigue.
Dippy is a shell command hook for Claude Code that automatically approves safe commands and blocks destructive ones to reduce permission fatigue. It solves the problem of constant approval prompts for basic operations like `ls` or `git status`, which disrupt developer flow, while still protecting against unsafe actions. Built on the Parable bash parser, it analyzes command complexity without external dependencies.
Developers using Claude Code who are frustrated by frequent permission prompts for non-destructive shell commands and want to maintain productivity without disabling safety entirely.
Developers choose Dippy because it intelligently distinguishes between safe and unsafe commands, auto-approving routine operations to save time while blocking dangerous ones. Its custom deny messages help guide Claude back on track, making AI-assisted development faster and more efficient.
🐤 Less permission fatigue, more momentum. Dippy knows what’s safe to run and keeps Claude on track when plans change.
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
Automatically approves non-destructive commands like `ls` and `git status`, reducing permission prompts and saving up to 40% development time, as highlighted in the README.
Handles intricate shell constructs such as pipelines and command substitution using the Parable parser, with over 14,000 tests ensuring accurate safety checks without external dependencies.
Allows configurable feedback when blocking commands, such as advising to use `trash` instead of `rm -rf`, preventing wasted turns and steering Claude back on track.
Supports global and per-project configuration files (e.g., `~/.dippy/config` and `.dippy`), enabling tailored safety rules for different environments, as documented in the wiki.
Only integrates with Claude Code, making it ineffective for users of other AI assistants or general shell environments, limiting its versatility.
Requires manual editing of Claude's settings.json or using hooks, which can be error-prone and less intuitive compared to plug-and-play tools.
Relies on predefined patterns for safe and destructive commands; may miss edge cases or newly introduced dangerous operations without frequent updates or custom configuration.