Menu
Por padrão, está selecionado o prompt que primeiro revisa a origem. Você pode mudar para um comando direto ou baixar uma cópia local.
Leia o SKILL.md e os arquivos complementares exibidos pelo SkillsMP antes de decidir se vai instalar.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Um comando direto ignora o prompt de revisão. Verifique a origem antes de executá-lo.
npx skills add https://github.com/jessedegans/breather --skill backO comando permanece em uma só linha. Role horizontalmente para revisá-lo antes de copiar.
Prefere uma cópia local? Baixe os arquivos disponíveis atualmente no SkillsMP.
Full break. Saves your mental context so you can stop without losing your place. Fully resets the fatigue clock.
Full break. Saves your mental context so you can stop without losing your place. Fully resets the fatigue clock.
End-of-session review. What you shipped, open threads, next steps.
Baseado na classificação ocupacional SOC
Exibindo SKILL.md
| name | back |
| description | Restore your mental context after a break. 5 seconds to get back in flow. |
| when_to_use | User says "I'm back", "back", "where was I", "what was I doing", "resume work". |
| allowed-tools | Bash Read |
The user is back from a break. Your job: get them oriented fast so they don't have to mentally reload.
Read the context snapshot from ${CLAUDE_PLUGIN_DATA:-~/.local/share/breather}/last-context.md. If the file doesn't exist, skip to step 3 and instead say: "No saved context from a previous pause. What are you picking up today?"
Read daily stats by running:
bash "${CLAUDE_PLUGIN_ROOT}/scripts/daily-stats.sh"
Welcome them back in one natural paragraph. Write it like you're talking to them, not filing a report. Reconstruct their mental state -- what was in their head when they stopped, what they were about to do next, and the one thing that matters most right now.
Good example:
You were making the indexing pipeline crash-resilient. All 7 tasks are done, tests pass. Next up: run the end-to-end test (
make run-deps-semdex && make fetch && make index), then write the ADR. The ADR is the last thing before this is merge-ready.
Bad example (too structured, requires parsing):
Working on: BAC-977 Resilient Blue/Green Indexing Pipeline Left off at: All 7 implementation tasks done on feature/bac-977 Next step: Test end-to-end, then ADR, then PR
The good version reads like a colleague catching you up in 10 seconds. The bad version reads like a JIRA ticket.
End with "Ready to pick up?" or similar.
Archive the context snapshot so calling /back twice doesn't show stale context:
mv "${CLAUDE_PLUGIN_DATA:-~/.local/share/breather}/last-context.md" "${CLAUDE_PLUGIN_DATA:-~/.local/share/breather}/last-context.md.used"
Do NOT ask how their break was, comment on how long they were gone, or add any preachy messaging. They're in work mode now. Respect that.