handoff

A session handoff skill for saving and restoring work context in docs/STATE.md. It records what is in progress, recent decisions, blockers, and postponed ideas so work can continue later.

In plain words
What is it for?
Pausing a coding session, ending work, resuming a project, reviewing active specifications, and recording blockers or decisions.
Why use it?
It reduces the need to reconstruct project context after a human or coding-agent session ends. It also helps identify the next concrete action when work resumes.

Skill for Claude CodeCodex

Install

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.

agentmods
npx agentmods add skills/igoruehara/spec-driven/handoff
Any agent
npx skills add igoruehara/spec-driven --skill handoff
Clone the repo
git clone --depth 1 https://github.com/igoruehara/spec-driven

Made for: Claude Code, Codex.

Per session 77 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 722 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00077 $0.00722
Opus 5 $0.00039 $0.00361
Sonnet 5 $0.00015 $0.00144
Haiku 4.5 $0.00008 $0.00072

Measured 2d ago against content hash 503d1833b27a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

handoff 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 2d 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.

template/.claude/skills/handoff/SKILL.md · 41 lines

How it starts

The opening of the file, as written. The whole thing — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Skill: Handoff de sessão (pause / resume)

Mantém a continuidade do projeto via docs/STATE.md — a memória de trabalho volátil (diferente do ADR, que é decisão durável). Detecte a intenção pelo pedido ou pergunte.

Modo PAUSE (pausar / encerrar)

Atualize docs/STATE.md com o estado real desta sessão:

  1. Em andamento / próximo passo — a feature/spec ativa e a próxima ação concreta (específica: "implementar AC-3 no adapter X", não "continuar a feature").
  2. Decisões recentes — o que foi decidido. Se for difícil de reverter, crie/atualize o ADR e linke; o STATE só resume.
  3. Bloqueios — o que trava e quem/como destrava.
  4. Ideias adiadas / todos — o que ficou de fora de propósito, com o gatilho para reconsiderar.
  5. Consolide docs/lessons.md (condicional — só se houve aprendizado): a captura das lições é inline durante o trabalho, não aqui. No handoff você revisa e poda: alguma lição duplicada? alguma já estabilizou → promova pro destino durável (ADR / glossary / CLAUDE.md / TESTING) e remova de lessons.md, deixando só o link em "Promovidas". Não force entradas vazias.
  6. Marque a data e o autor. Se houver SPEC_DEVIATION aberto, registre como bloqueio.
  7. Ofereça um commit (docs: handoff — atualiza STATE) se for repositório git.

Seja conciso e acionável. O STATE é para alguém (ou um agente) retomar frio amanhã.

Modo RESUME (retomar)

  1. Leia docs/STATE.md e a spec.md/tasks.md da feature ativa citada nele.
  2. Se houver MCP conectado e relevante (Jira/Confluence), e a trava de conta estiver validada, atualize o contexto com o que mudou lá fora.
  3. Resuma onde paramos em poucas linhas: feature ativa, último passo, bloqueios abertos.
  4. Proponha o próximo passo (o "Em andamento / próximo passo" do STATE) e confirme com o usuário antes de executar.

Regras

  • STATE.md é volátil; ADR é durável. Não escreva decisão estrutural só no STATE.
  • lessons.md é staging: no handoff só consolida/poda — a captura é inline durante o trabalho. Lição que estabilizou tem que sair de lá; se está só crescendo, a poda falhou.
  • Não invente progresso: relate fielmente o que foi feito, o que falta e o que está bloqueado.

Read the full file on GitHub · 41 lines

Changes

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.

  1. 2d ago First seen · 41 lines · 77 tokens per session scan A 503d1833b27a

Subscribe to this mod's changes

handoff is a skill published in the GitHub repository igoruehara/spec-driven (213 stars, last pushed 1mo ago), licensed MIT. It adds 77 tokens to every session and 722 once invoked, about $0.0004 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 2026-08-30.