Agent skill · data analytics · aradotso
codeburn-claude-cost-dashboard
Interactive TUI dashboard for visualizing Claude Code token usage, costs, and task breakdowns by project, model, and activity type.
Why this skill is useful
Adds specific commands and local file analysis for tracking Claude Code token usage and costs that the AI wouldn't reliably generate on its own.
What it needs
Requires node installed locally. About 5k tokens when loaded. Last updated 2026-07-09. 70 stars on the source repository.
What this skill does
CodeBurn Skill by ara.so — Daily 2026 Skills collection. CodeBurn is an interactive TUI dashboard that reads Claude Code session transcripts directly from disk (~/.claude/projects/) and visualizes token usage, costs, and task breakdowns by project, model, activity type, MCP server, and tool. No API keys, no proxy — pure local file analysis. Install Requirements: Node.js 20+, Claude Code installed (session data at ~/.claude/projects/) Key Commands TUI Keyboard Navigation Key Action ----- -------- ← → arrow keys Switch between Today / 7 Days / Month 1 2 3 Period shortcuts q Quit What CodeBurn Tracks 13 Task Categories (deterministic, no LLM calls) Category Trigger Pattern --- --- Coding Edit, Write tools Debugging Error/fix keywords + tool usage Feature Dev "add", "create", "implement" keywords Refactoring "refactor", "rename", "simplify" Testing pytest, vitest, jest in Bash Exploration Read, Grep, WebSearch without edits Planning EnterPlanMode, TaskCreate tools Delegation Agent tool spawns Git Ops git push/commit/merge in Bash Build/Deploy npm build, docker, pm2 Brainstorming "brainstorm", "what if", "design" Conversation No tools, pure text exchange General Skill tool, uncategorized One-Shot Rate For edit-heavy categories, CodeBurn detects Edit → Bash → Edit retry cycles. The 1-shot rate shows what percentage of edit turns succeeded without retries. 90% means the AI got it right first try 9/10 times. Pricing Fetched from LiteLLM model prices, auto-cached 24h at ~/.cache/codeburn/. Covers input, output, cache write, cache read, and web search costs. Falls back to hardcoded prices if fetch fails. Data Source Claude Code stores session transcripts as JSONL: Each assistant entry contains: model name, token usage (input/output/cacheread/cachewrite), tooluse blocks, timestamps. CodeBurn deduplicates by API message ID to prevent double-counting across sessions. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills aradotso/codeburn-claude-cost-dashboard