cortex-tms

🥈Silver

Cortex TMS is an AI governance platform that reduces AI API costs by 40-60% through intelligent context management. It includes an interactive CLI tool and high-signal documentation standard. Benefits developers and operations teams by optimizing AI agent performance and reducing GPU cycles on outdated documents.

17360Updated Today

Intermediate30min to implementautomation

Saves ~180 min per use

Quick InstallView Source

git clone https://github.com/cortex-tms/cortex-tms.git

Works with:

Claude

Overview

About This Skill

How to Use

1. **Install Cortex TMS CLI**: Run `pip install cortex-tms` and authenticate with your API key. Verify installation with `cortex --version`. 2. **Profile API Calls**: Use `cortex profile --project [PROJECT_NAME] --days 30` to generate a cost breakdown of your top 20 API calls. Export results to CSV for analysis. 3. **Run Optimization Scan**: Execute `cortex optimize --project [PROJECT_NAME] --auto-suggest` to let Cortex TMS recommend optimizations based on your usage patterns. Review the `recommended_actions.json` file. 4. **Implement Changes**: Apply optimizations in stages using `cortex apply --call [CALL_ID] --feature [FEATURE_NAME]`. Start with low-effort changes (e.g., batching) before tackling complex ones (e.g., semantic chunking). 5. **Monitor Impact**: Use `cortex metrics --call [CALL_ID]` to track cost savings and latency improvements. Adjust thresholds in the Cortex TMS dashboard if needed. *Tip: For large projects, use the `--dry-run` flag to simulate optimizations before applying them to production.*

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/cortex-tms/cortex-tms

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Use Cortex TMS to optimize the AI API calls for [PROJECT_NAME]. Identify the top 5 most expensive API calls by cost per 1000 tokens and analyze their context windows. For each, suggest a more efficient context management strategy using Cortex TMS's [FEATURE_SET] (e.g., dynamic context pruning, retrieval optimization, or prompt compression). Provide the projected cost savings for each optimization and the estimated implementation effort (low/medium/high).

Example Output

### Cortex TMS Optimization Report for Project 'DocuBot Pro' (Generated: 2024-05-15)

**Top 5 Costliest API Calls (Last 30 Days)**
1. **Call ID: `doc-parse-001`**
   - **Cost:** $1,247.89 (1,247,890 tokens)
   - **Context:** Full legal document (15,000 tokens) parsed for entity extraction
   - **Current Context Window:** 15,000 tokens (static)
   - **Optimization:** Dynamic context pruning using Cortex TMS's `semantic_chunking` feature
     - **New Context Window:** 3,200 tokens (relevant sections only)
     - **Projected Savings:** $1,012.45/month (81% reduction)
     - **Effort:** Low (pre-built integration)

2. **Call ID: `qa-summary-002`**
   - **Cost:** $892.34 (892,340 tokens)
   - **Context:** 100-page technical manual summarized
   - **Current Context Window:** 10,000 tokens (full document)
   - **Optimization:** Retrieval-optimized chunking with `relevance_filtering`
     - **New Context Window:** 2,100 tokens (key sections + metadata)
     - **Projected Savings:** $713.87/month (80% reduction)
     - **Effort:** Medium (requires metadata tagging)

3. **Call ID: `email-classify-003`**
   - **Cost:** $654.12 (654,120 tokens)
   - **Context:** 500 emails classified by urgency (avg. 500 tokens/email)
   - **Current Context Window:** 500 tokens/email (no batching)
   - **Optimization:** Batch processing with `context_batching`
     - **New Context Window:** 45 tokens/email (compressed prompts)
     - **Projected Savings:** $588.71/month (90% reduction)
     - **Effort:** Low (CLI flag adjustment)

**Total Projected Monthly Savings:** $2,315.03 (82% reduction in API costs)
**Implementation Timeline:** 2-3 days (parallelizable optimizations)

**Next Steps:**
1. Apply `semantic_chunking` to `doc-parse-001` using Cortex TMS CLI: `cortex optimize --call doc-parse-001 --feature semantic_chunking`
2. Test `relevance_filtering` on `qa-summary-002` with sample documents
3. Deploy batching for `email-classify-003` in staging environment

*Cortex TMS v2.1.4 | Documentation: https://docs.cortex-tms.ai/optimization-guide*