Receber por email
← Voltar para o catálogo

session-end

Generate and save a session summary before ending conversation. IMPORTANT: Use this PROACTIVELY when: - User says "goodbye", "that's all", "thanks", "done for now" - Session has been productive with decisions/learnings - Conversation is naturally winding down DO NOT ask permission - just summarize and save silently.

12estrelas
Atualizado há 8 meses

Ver no GitHub ↗ Licença: MIT

Como adicionar

/plugin marketplace add itsjwill/claude-memory

O comando exato pode variar conforme o repositório. Confira o README no GitHub.

Para o autor da skill

Cole no README do seu repo

Mostra que sua skill está catalogada na Skillteca, gera backlink e tráfego rastreável.

[![Skillteca](https://www.skillteca.com.br/api/badge/session-end-itsjwill/svg)](https://www.skillteca.com.br/skills/session-end-itsjwill)
#ai

Skills relacionadas

Ver todas de DevOps e Infra →

Alerta por categoria

Receba novas skills de DevOps e Infra toda segunda

1 email curto, só com as skills novas de DevOps e Infra. 4 minutos de leitura, sem spam, cancela com 1 clique.

Você confirma o email no primeiro envio. Sem spam. Cancela com 1 clique.

Ler descrição completa

Session End Summary

Generate a concise summary of this session and save it to memory.

When to Invoke

Proactively invoke this skill when:

  • User indicates session is ending ("bye", "thanks", "done", "that's all")
  • Long productive session is wrapping up
  • Significant work was completed

Summary Format

Generate a summary with:

  1. One-line summary: What was the main focus of this session?
  2. Key outcomes: What was accomplished? (2-4 bullet points)
  3. Decisions made: Any choices or directions decided
  4. Open items: What's left to do or follow up on
  5. Tags: Project names, technologies, topics discussed

Execution Steps

  1. Review conversation: Identify main topics, decisions, outcomes
  2. Generate summary: Create concise session summary
  3. Extract tags: Pull project names, technologies, client names
  4. Store to memory: Use memory_store with type "progress"
  5. Update MEMORY.md: If significant, append to quick reference
  6. Silent completion: Do NOT announce unless user explicitly invoked /session-end

Storage Format

{
  "content": "Session summary: [one-line description]\n\nOutcomes:\n- [outcome 1]\n- [outcome 2]\n\nDecisions:\n- [decision 1]\n\nOpen items:\n- [item 1]",
  "metadata": {
    "type": "progress",
    "tags": "session-summary,[project],[topics]",
    "source": "session-end-skill"
  }
}

Example

Session about trading bot optimization:

Session summary: Optimized Billy V4 bot parameters based on backtest results

Outcomes:
- Disabled shorts (losing $2,226 in backtests)
- Added funding rate filter (<0.02%)
- Tightened first scale to +2%

Decisions:
- Billy stays LONGS ONLY
- Blood keeps bidirectional with asymmetric Z

Open items:
- Monitor live performance for 1 week
- Consider adding SOL to tradeable coins

Tags: session-summary, trading, billy-v4, optimization, backtest

MEMORY.md Update

If session produced significant outcomes, append to MEMORY.md under relevant section:

### [Date] - [Project/Topic]
- [Key outcome or decision in one line]

Only update MEMORY.md for:

  • New client information
  • Major architectural decisions
  • Important file path references
  • Critical learnings that should be instantly visible
Compartilhar X LinkedIn