Free skills
The free, one-shot versions of Cost
Open-source Claude Agent Skills for LLM cost and FinOps. Each one looks at a single prompt, run, or usage export and hands you a concrete dollar number. No API key, no account - they run inside your Claude session. Cost does the same analysis continuously, across all your production traffic.
Install all of them in Claude Code:
/plugin marketplace add botzone-ai/llm-cost-skills
/plugin install llm-cost-skills@llm-cost-skillsspend-teardown
Parse a provider usage CSV into a spend breakdown by model and input/output split, with a ranked savings list.
npx skills add botzone-ai/llm-cost-skills/skills/spend-teardownprompt-caching-advisor
Find the static, repeated segments of a prompt and project the input-cost cut from prompt caching, per provider.
npx skills add botzone-ai/llm-cost-skills/skills/prompt-caching-advisorcontext-bloat-auditor
Read a prompt or agent run, flag duplicated and oversized context, and price the wasted tokens per call and per month.
npx skills add botzone-ai/llm-cost-skills/skills/context-bloat-auditormodel-downgrade-advisor
Judge whether a task can run on a cheaper model, project the saving, and specify how to verify that quality holds first.
npx skills add botzone-ai/llm-cost-skills/skills/model-downgrade-advisorbatch-api-advisor
Spot workloads that can move to the Batch API for roughly half price, and estimate the monthly saving.
npx skills add botzone-ai/llm-cost-skills/skills/batch-api-advisorOne prompt is a snapshot. Your bill is continuous.
These skills analyse one artifact at a time. Cost wraps your Anthropic, OpenAI, and Gemini clients and attributes every euro to a route or feature across all your traffic, with monitoring and alerts. Free under €2,000 a month of model spend, no card.
Track your spend continuously