Cache Tax — Development skill for Claude Code
Keep Claude Code’s prompt cache warm during breaks and show the estimated cost before a cold send.
How to install Cache Tax
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open karanb192/cache-tax and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Cache Tax does
Keep Claude Code’s prompt cache warm during breaks and show the estimated cost before a cold send.
Alternatives in Development
- Mono Color Skill — One-ink editorial print image skill — warm paper, halftone photography, active negative space, and restrained 3k ★
- Cost Xray — See what Claude Code and Codex actually send to the API — and what each part costs 586 ★
- Mega Plan — Generate a mega-plan for project-level multi-feature orchestration 123 ★
README
cache-tax
**Keep Claude Code's prompt cache warm during breaks.**
On Fable 5.1, a one-hour cache write costs **80x a cache read per token**: [$20 versus $0.25 per million tokens](https://platform.claude.com/docs/en/about-claude/pricing). Run `/keepwarm` before stepping away. If you return cold, the guard stops your send once with the estimated rewrite price.
**[Install](#install)** · [How it works](#how-it-works) · [Costs and limits](#cost-and-the-plan-limit-question) · [Commands](#commands)

A two-line recap in a 330k-token session triggered a **$6.61 estimate**. After resending, the reported cache write was **$6.28**. Recorded on 2.1.1; current wording says “up to” the estimated token count. These are API-equivalent costs, not extra subscription charges.
Install
Needs early-access function hooks, a one-hour cache and Claude Code left running. Pings cost tokens, including uncapped output.
claude plugin marketplace add karanb192/claude-code-mods
claude plugin install cache-tax@claude-code-mods
Start Claude Code with function hooks enabled:
CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1 claude
In a warm session, run `/keepwarm` to arm six hours. Use `/keepwarm 90m` for a shorter window, or `/keepwarm off` to stop.
**Check the cache lifetime.** Included subscription usage defaults to one hour for the main conversation; API-key and usage-credit requests default to five minutes. Forks belong to a separate bucket. The [documented controls](https://code.claude.com/docs/en/prompt-caching#choose-the-ttl-yourself) are `promptCacheTtl` for the main conversation and `subagentPromptCacheTtl` for requests outside it; each accepts `"1h"`. A 50-minute timer does not protect a five-minute cache. See [Limits](#limits) for the evidence boundaries.
**No function hooks yet?** The [hook version](https://github.com/karanb192/claude-code-hooks/tree/main/plugins/cach
Related Skills
Claude Code Coffee
Keep your Claude Code prompt cache warm during breaks. Saves up to 92% on cache rebuild costs.
Wirescope
A transparent analytical proxy for the Claude Code & Codex CLIs — see the real per-turn cost, prompt-cache war
Cc Cache Keepalive
Keep the most recently used Claude Code prompt cache warm without polluting the original session.
Offline Telemetry Script
Offline analyzer for Claude Code session transcripts (~/.claude/projects/.jsonl). Pure-stdlib Python, no API c
Token Cost
Show token usage and estimated cost for this project
Herdr Auto Warm Cache
A herdr plugin that keeps an idle agent's prompt cache warm before the one-hour TTL expires — 20x cheaper than
Related Agents
Worker Apply Opus
Resumable implementation worker for non-SDD work units. Applies multi-file changes from a handoff, absorbs fol
Stats Analyzer
Analyzes ~/.claude/stats-cache.json and history.jsonl to produce a concise usage report with model token break
Usage Reporter
Reads real Claude Code usage from ~/.claude/stats-cache.json. Reports daily/weekly token usage by model, cache