Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add luanpdd/kit-mcp --skill cost-trackinggit clone --depth 1 https://github.com/luanpdd/kit-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/luanpdd/kit-mcp/cost-tracking)<a href="https://agentmods.dev/skills/luanpdd/kit-mcp/cost-tracking"><img src="https://agentmods.dev/badge/skills/luanpdd/kit-mcp/cost-tracking.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00052 | $0.03788 |
| Opus 5 | $0.00026 | $0.01894 |
| Sonnet 5 | $0.00010 | $0.00758 |
| Haiku 4.5 | $0.00005 | $0.00379 |
Grade A, and why
cost-tracking scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 282 lines — stays where its author put it; the contents beside it link to each section on GitHub.
kit-mcp — Cost Tracking (Claude Code USD + tokens)
Quando usar
LLM carrega esta skill quando o user pergunta sobre dinheiro/tokens gastos com Claude Code. Trigger phrases:
- "custo", "cost", "USD", "dólares", "gasto", "gastei", "spent"
- "tokens", "token usage", "consumo de tokens"
- "pricing", "billing", "fatura"
- "budget" (no sentido financeiro, NÃO error budget)
- "ROI", "retorno", "vale a pena essa fase"
- "ccusage", "rastrear custo Claude"
- "quanto a sessão custou", "custo do dia", "custo da fase"
- "estimar custo de prompt", "estimate cost"
- "blocks 5h", "janela de 5 horas"
Disambiguation — leia ANTES de chamar
Esta skill é sobre USD/tokens gastos pelo user com a API Claude. Três skills/comandos parecem similares no nome mas resolvem problemas diferentes — não confundir:
| Skill / Tool | Mede o quê | Quando usar |
|---|---|---|
cost-tracking (esta) |
USD reais gastos com Claude API (snapshot LiteLLM pricing) | "Quanto gastei hoje?" / "Custo da fase 172?" |
burn-rate-status (cmd) |
Error budget SLO queimando (ratio de erros vs target de disponibilidade) | "Meu SLO está estourando?" / "Page on-call?" |
risk-budget (skill SRE) |
Risco SRE acumulado (mudanças vs budget de error tolerance) | "Posso fazer release hoje sem queimar budget?" |
metrics-snapshot (tool) |
Golden signals do próprio kit-mcp server (latency p99, errors, traffic) | "MCP server está saudável?" |
Regra mnemônica: se a pergunta tem "$", "USD", "tokens", "fatura" → cost-tracking. Se tem "SLO", "5xx", "error rate", "on-call" → burn-rate-status ou risk-budget.
Regras absolutas
- Modelo desconhecido NUNCA retorna
$0silencioso — entra emunknown_models[]comusd: null. Mostra esses modelos ao user para que ele saiba o que NÃO foi precificado. - Pricing tem staleness —
pricing_staleness_daysno output. Se > 30d, hápricing_warning. Refresh manual vianpm run regen-pricing(M5) ou GH Action semanal. - Naming kebab-case —
cost-today, NÃOcost_today. Consistente commetrics-snapshot,reverse-sync. cost-estimateé heurística — chars/4 ± 30%, sem tokenizer real na v1.37.0. Odisclaimerno output explicita isso. NÃO citar valor exato para o user; sempre apresentar[low, high]ou avisar do range.cost-phaseé correlação, não medição direta — usa janela temporal inferida (mtime de SPEC.md + STATE.md.completed_at + git log). O campocorrelation_confidence(high/medium/low/unknown) gradua a confiança. NUNCA reportarcost-phasecomo número exato sem dizer a confidence.persist:trueé opt-in — default off. Quando ligado, grava.planning/costs/<ts>.json(gitignored). Útil para histórico, ruim para CI ruidoso.refresh_pricing:trueé opt-in com side-effects — fetch http models.dev + cache em~/.kit-mcp/pricing-cache.jsonTTL 24h. NÃO ligar por padrão (latência + risco rede + pollution de cache).- Timezone default é UTC — paridade com
ccusage. Override viatzse o user quer "meu dia" local (ex:tz: "America/Sao_Paulo"). - Blocks são janelas de 5h sliding com gap-detection — gap > 5h entre entries inicia novo bloco. Útil para ver "sessão de trabalho A" vs "sessão B" do mesmo dia.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 282 lines · 52 tokens per session scan A 71ca8046e9f0
cost-tracking is a skill published in the GitHub repository luanpdd/kit-mcp (1 stars, last pushed yesterday), licensed MIT. It adds 52 tokens to every session and 3,788 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
sector-rotation
An analysis framework for comparing industries in the Chinese A-share stock market, using business conditions, price momentum, valuation, and money flows. It produces rankings and higher- or lower-allocation suggestions.
strategy-pivot-designer
Detect backtest iteration stagnation and generate structurally different strategy pivot proposals when parameter tuning reaches a local optimum.
twitter-reader
Read Twitter/X for financial research using opencli (read-only). Use this skill whenever the user wants to read their Twitter feed, search for financial tweets, view bookmarks, look up user profiles, or gather market sentiment from Twitter/X. Triggers include: "check my feed", "search Twitter for", "show my…
chenhao-limit-up
A framework for judging Chinese A-share stocks that have reached the daily price-rise limit, using market mood, sector leadership, and trading momentum.
trading-risk-gate
Unified pre-trade safety gate: Ruin check (Law #1), ergodicity audit, and win-rate dominance validation. Absorbs: ergodicity-check, law-of-ruin, win-rate-dominance.
vectorbt
High-performance vectorized backtesting with parameter optimization, portfolio simulation, and rich performance metrics.