Stats
Actions
Tags
From dev-team-kit-fv
Audits fixed and dynamic context overhead in the current session — scans CLAUDE.md, agents, MCP servers, rules, skills, and history — estimates tokens per component, reports budget usage and headroom, and issues alerts for oversized components or overflow risk.
How this command is triggered — by the user, by Claude, or both
Slash command
/dev-team-kit-fv:context-budgetThe summary Claude sees in its command listing — used to decide when to auto-load this command
# /context-budget Audita o **overhead de contexto fixo e dinâmico** da sessão atual. **Diferença do `/savings`:** savings mostra tokens economizados pelo kit em runtime; context-budget mostra o que já está carregado no context window antes de qualquer completion. ## O que faz 1. Varre componentes fixos: `CLAUDE.md` (global + projeto), `GLOBAL.md`, `agents/*.md`, descrições de MCP ativos 2. Varre componentes dinâmicos: rules ativadas nesta sessão, skills invocadas, histórico acumulado 3. Estima tokens por componente (bytes ÷ 4, aproximação BPE) 4. Reporta tabela com % do budget e headroo...
Audita o overhead de contexto fixo e dinâmico da sessão atual.
Diferença do /savings: savings mostra tokens economizados pelo kit em runtime; context-budget mostra o que já está carregado no context window antes de qualquer completion.
CLAUDE.md (global + projeto), GLOBAL.md, agents/*.md, descrições de MCP ativos/swarm ou /loop --parallel — garantir headroom suficiente.bot/ — verificar o que foi instalado/context-budget
Skill carregada: dev-team-kit-fv:49-context-budget
## Context Budget — [repo] — [data]
### Fixo (~14.400 tokens, 72% do subtotal)
- CLAUDE.md global: ~2.400 tokens
- agents/ (16): ~8.000 tokens
- MCP servers (N): ~X tokens
...
### Dinâmico (esta sessão)
- rules/common/: ~3.200 tokens
- skills invocadas: ~N tokens
- histórico: ~N tokens
### Status: ✅ Saudável | ⚠️ Atenção | 🚨 Overflow iminente
### Headroom: X tokens (Y%)
### Recomendações (se houver)
...
npx claudepluginhub felvieira/claude-skills-fv --plugin dev-team-kit-fv/cc-budgetCalculates remaining context window after fixed overhead from CLAUDE.md, rules, skills, and MCP schemas. Shows budget summary table and top consumers. Also supports audit, optimize, anchor-plan, simulate.
/contexteRuns built-in /context command to fetch real token metrics, analyzes usage patterns for inefficiencies, and provides tailored optimization recommendations.
/context-budgetAnalyzes context window usage across agents, skills, MCP servers, and rules to find optimization opportunities, producing a Context Budget Report with recommendations.
/fire-costTracks context window consumption and estimates session costs, recommending strategic compaction before hitting limits. Supports --detail, --warn, and --budget flags.
/optimize-contextAnalyzes context window usage against MECW principles, providing token breakdowns, utilization status, and optimization recommendations. Accepts optional path and --detailed flag.
/claude-pulseDisplays real-time session dashboard with cost analysis, token breakdown, tool performance, context usage, and optimization tips. Also supports configure, stats, setup subcommands.