菜单
Analyze Claude Code token consumption and estimated costs across projects and sessions. Use when user asks about token usage, costs, spending, billing, how much they've used, or wants to optimize their Claude Code usage patterns.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use this skill when the user wants a browser-native HTML presentation, web deck, single-file slides, horizontal swipe deck, keynote-style web page, investor/demo-day HTML deck, or a shareable presentation that opens in a browser. Also use when the user asks for an editorial deck, clean grid deck, presentation microsite, slide landing page, or HTML slides. Do not use this for .pptx files; use the pptx skill for PowerPoint input or output.
Use when the user asks to inspect AdClaw Host operators, customers, containers, sandbox statuses, stuck runtime alerts, backups, runtime reconcile, or safe admin actions.
Use for AdClaw Host AI quota accounting, hosted key provisioning, limit UX, and secret redaction reviews.
Analyze prompts for constraint complexity, audit failure risks, and generate optimized rewrites for Claude and GPT. Based on "How LLMs Follow Instructions" (Rocchetti & Ferrara, 2026) constraint taxonomy research. Use when reviewing prompt files, optimizing prompt bases, or auditing instruction quality. Trigger phrases — "analyze prompt", "prompt audit", "optimize prompts", "constraint analysis", "prompt review", "check prompts".
Install and operate a portable OpenAI Symphony + Linear orchestration setup in any Git repository. Use when the user wants a one-command local Symphony runner, a reusable WORKFLOW template, background launch scripts, or a Codex wrapper that restarts Symphony automatically when reopening the CLI.
Parallel multi-agent code review using Agent Teams with 4 specialized reviewers. Spawns a coordinated team of security, performance, test coverage, and code quality agents. Teammates can share findings with each other for cross-domain insights. Produces unified report with severity-ranked findings saved to /output/. Requires Claude Code with Agent Teams support (TeamCreate, TaskCreate, SendMessage). Use when asked to: - "review code", "code review", "check my changes", "review PR" - "security review", "performance review", "test coverage check" - "full code review", "multi-agent review", "parallel review" Examples: - `/code-review-team` -- review uncommitted changes (git diff) - `/code-review-team 42` -- review PR #42 - `/code-review-team --scope=security` -- security-only review of uncommitted changes - `/code-review-team 42 --scope=performance,tests` -- specific reviewers for PR
| name | token-usage |
| description | Analyze Claude Code token consumption and estimated costs across projects and sessions. Use when user asks about token usage, costs, spending, billing, how much they've used, or wants to optimize their Claude Code usage patterns. |
Analyze token consumption patterns and estimate costs from Claude Code session files (~/.claude/projects/).
Run the analyzer with a time range argument:
python3 .claude/skills/token-usage/scripts/analyze.py $ARGUMENTS
| Argument | Example | Description |
|---|---|---|
| (empty) | All time | |
today | today | Since midnight UTC |
week | week | Last 7 days |
month | month | Last 30 days |
| N (number) | 3 | Last N days |
| Date | 2026-04-01 | Since that date |
| Datetime | 2026-04-01 11:00 | Since that datetime |
--compare — show current period vs previous period of same lengthExample: --compare week shows this week vs last week.
The script prints a summary to stdout and saves a detailed report to:
~/.claude/token-usage-reports/token_report.md
Prices are per million tokens, detected automatically per model:
| Model | Input | Cache Create | Cache Read | Output |
|---|---|---|---|---|
| Opus | $15.00 | $18.75 | $1.50 | $75.00 |
| Sonnet | $3.00 | $3.75 | $0.30 | $15.00 |
| Haiku | $0.80 | $1.00 | $0.08 | $4.00 |
The analyzer reads the model field from each assistant message and applies correct pricing automatically.
--compare)~/.claude/projects/)