router-stats
The router-stats Claude Code skill reads usage statistics from a local JSON file and displays Claude Router metrics including query distribution across model tiers, exception handling details, and estimated cost savings compared to always using the most capable model. Use this skill to track how effectively the Claude Router is optimizing query routing and to visualize accumulated savings from routing simpler requests to faster, cheaper models.
git clone --depth 1 https://github.com/aiskillstore/marketplace /tmp/router-stats && cp -r /tmp/router-stats/skills/0xrdan/router-stats ~/.claude/skills/router-statsSKILL.md
# Router Stats
Display usage statistics and estimated cost savings from Claude Router.
## Instructions
Read the stats file at `~/.claude/router-stats.json` and present the data in a clear, formatted way.
## Data Format
The stats file contains (v1.2 schema):
```json
{
"version": "1.2",
"total_queries": 100,
"routes": {"fast": 30, "standard": 50, "deep": 10, "orchestrated": 10},
"exceptions": {"router_meta": 15, "slash_commands": 0},
"tool_intensive_queries": 25,
"orchestrated_queries": 10,
"estimated_savings": 12.50,
"delegation_savings": 2.50,
"sessions": [
{
"date": "2026-01-03",
"queries": 25,
"routes": {"fast": 8, "standard": 12, "deep": 2, "orchestrated": 3},
"savings": 3.20
}
],
"last_updated": "2026-01-03T15:30:00"
}
```
## Output Format
Present the stats like this:
```
╔═══════════════════════════════════════════════════╗
║ Claude Router Statistics ║
╚═══════════════════════════════════════════════════╝
📊 All Time
───────────────────────────────────────────────────
Total Queries Routed: 100
Route Distribution:
Fast (Haiku): 30 (30%) ████████░░░░░░░░░░░░
Standard (Sonnet): 50 (50%) ██████████████░░░░░░
Deep (Opus): 10 (10%) ████░░░░░░░░░░░░░░░░
Orchestrated: 10 (10%) ████░░░░░░░░░░░░░░░░
🔧 Tool-Aware Routing
───────────────────────────────────────────────────
Tool-Intensive Queries: 25 (25%)
Orchestrated Queries: 10 (10%)
⚡ Exceptions (handled by Opus despite classification)
───────────────────────────────────────────────────
Router Meta-Queries: 15 (queries about the router itself)
Total Exceptions: 15
💰 Cost Savings
───────────────────────────────────────────────────
Estimated Savings: $12.50 (compared to always using Opus)
Delegation Savings: $2.50 (from hybrid delegation)
Total Savings: $15.00
📅 Today (2026-01-03)
───────────────────────────────────────────────────
Queries: 25
Savings: $3.20
Route Distribution:
Fast: 8 | Standard: 12 | Deep: 2 | Orchestrated: 3
```
## Steps
1. Use the Read tool to read `~/.claude/router-stats.json`
2. If the file doesn't exist, inform the user that no stats are available yet
3. Calculate percentages for route distribution
4. Display exception counts if present (router_meta queries are handled by Opus despite classification)
5. Format and display the statistics
6. Include the savings comparison explanation
## Notes
- Savings are calculated assuming Opus would have been used for all queries
- Cost estimates use: Haiku 4.5 $1/$5, Sonnet 4.5 $3/$15, Opus 4.5 $5/$25 per 1M tokens
- Average query estimated at 1K input + 2K output tokens
- **Exceptions**: Queries about the router itself are classified but handled by Opus (per CLAUDE.md rules). This is intentional - users discussing the router get the most capable model while still seeing what the classifier decided.Let Codex use Google Antigravity CLI (agy) for repository exploration, bounded feature work, and project-scale implementation. Use when a Codex task benefits from delegated repository work while Codex retains diff review, verification, repair, and delivery assurance.
Use when replying. Refresh current date and time.
Create polished, validated architecture, workflow, sequence, data-flow, and lifecycle/state diagrams as explorable standalone HTML with inline SVG, dark/light themes, optional trace motion, and PNG/JPEG/WebP/SVG/WebM export. Accept plain-language requirements or pasted Mermaid flowchart, sequenceDiagram, and stateDiagram input; inspect repository evidence when the diagram must reflect real code. Use when the user asks to visualize system architecture, infrastructure, cloud/security/network topology, technical workflows, API call sequences, request lifecycles, data pipelines, ETL/ELT, data lineage, state machines, or to convert/beautify Mermaid.
Design systems, plan implementations, review architecture decisions - Use when you need to plan a complex feature, design system architecture, or make high-level technical decisions.
Use when automating iOS Simulator interactions, capturing screenshots/video, or inspecting accessibility via AXe CLI.
Use when working with Canton Network participants, DAML smart contracts, Splice applications, or debugging LF version and package ID issues.
>
HSK4級レベルから流暢さを目指す学習者向け。中国語表現の使用場面・自然さを分析し、作文を「ネイティブらしい流暢な表現」に改善。bilibili等のコンテンツ理解とネイティブとの会話をサポート。実際の用例をWeb検索で提示