claude-permissions-optimizer

To reduce Bash permission prompts in Claude Code, analyze your session history to identify frequently used safe commands and generate allowlist patterns for your settings.json. This prevents safe commands from repeatedly requesting approval during development.

You can create a Bash allowlist from Claude Code session logs by scanning transcripts to extract tool usage, normalizing commands into stable patterns, and filtering them by safety classification. Safe patterns are then appended to your project or user settings.

Yes, allowlist optimization works with other coding agent transcripts. The extraction scripts analyze recent transcripts across Claude Code sessions and other agents to normalize Bash commands and surface safe patterns for your settings.

Bash commands are classified into safe, review, and dangerous categories during permission optimization. Commands are normalized into stable patterns, filtered to remove already-covered rules and low-frequency noise, leaving only safe-to-allow patterns for review.

Yes, you can automatically apply safe Bash patterns to Claude Code settings.json. The optimizer optionally appends approved rules to user or project settings with JSON validation and write verification to ensure the configuration is safely updated.

A limitation of generating Bash allowlists from session history is that dangerous commands may still appear in transcripts and require manual review. The tool filters low-frequency noise and already-covered commands, but users must verify safe patterns before applying them to settings.