By b-open-io
Permission analytics TUI for analyzing Claude Code permission usage patterns
A TUI tool for analyzing and managing Claude Code permission usage. Parses session logs to show which tools are being called, how often they're approved vs denied, and lets you apply permissions to your settings files with a live diff preview.

go install github.com/b-open-io/claude-perms/cmd/perms@latest
Or build from source:
git clone https://github.com/b-open-io/claude-perms.git
cd claude-perms
go build -o bin/perms ./cmd/perms/
perms
Frequency — All tool permissions ranked by usage, with Allow/Deny counts parsed from session logs. Expand groups to see individual variants (e.g., Bash(curl:*), Bash(git:*)). Press Enter on any permission to open the apply modal.
Matrix — Agents and their actual tool usage across sessions. Shows which permissions each agent type requests. Select an agent to multi-select permissions and batch-apply them.
Help — Keyboard shortcuts reference.
When you apply a permission, the modal shows a live diff preview of the exact settings file that will be edited, with line numbers and colored +/- lines. After applying, a toast notification confirms the file and line that was written.
Permissions are written to:
~/.claude/settings.local.json<project>/.claude/settings.local.json| Key | Action |
|---|---|
j/k | Navigate |
Enter | Expand group / Open apply modal |
Tab | Switch views |
/ | Filter permissions |
Esc | Close modal / Clear filter |
q | Quit |
| Key | Action |
|---|---|
Space | Toggle permission selection |
A | Apply selected permissions |
j/k | Navigate |
Esc | Close |
Parses JSONL session logs from ~/.claude/projects/ to extract tool_use events and correlate them with tool_result responses. User denials are detected by checking for is_error: true with content containing "rejected" — command failures (exit codes, etc.) are not counted as denials.
Results are cached in ~/.claude/perms-cache.json for fast subsequent launches.
MIT
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub b-open-io/claude-plugins --plugin claude-permsDevelopment agents, skills, hooks, and commands for Claude Code workflows
1Sat ecosystem tools for BSV — unified indexing API (1sat-stack), ordinals minting and marketplace (list/buy/cancel), BSV21 token operations, wallet setup (BRC-100), transaction building, time locks, sweep/import, OpNS names, dApp wallet connection, CLI tool, media extraction, and MCP server for wallet-desktop browser automation.
Core BSV blockchain operations including standards reference (BRCs, BitCom, tokens), key derivation (Type42, BIP32, BAP), ORDFS content gateway, script templates, message signing, wallet operations, identity management, and JungleBus real-time blockchain streaming.
Gemini 3.1 Pro skills powered by Nano Banana 2, Gemini 3 Flash, and Veo 3.1 - image generation with 169 art styles including pop culture icons and video game aesthetics, video generation (text-to-video, image-to-video) with native audio, interactive style browser with tile regeneration, pixel avatars, team group photos, section dividers, platform icons (favicon, iOS, Android, PWA, desktop), image optimization with sharp, text generation, upscaling, editing, SVG creation, segmentation, visual workflow planning with tldraw infinite canvas, and dynamic Gemini API docs via llms.txt.
ClawNet bot templates and skills for AI agent development on the ClawNet platform
OTel + Hooks analytics for Claude Code. Run /init-claudalytics to configure telemetry and hooks capture for this project.
Tracks Claude Code session usage (tokens, costs, skills, MCP, subagents) and sends data to a self-hosted dashboard.
Analyze Claude Code session token usage, costs, trends, and anomalies
Session intelligence for Claude Code — find why your sessions fail and fix them
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Frontend design skill for UI/UX implementation