Menu
Le prompt qui vérifie d'abord la source est sélectionné par défaut. Vous pouvez passer à une commande directe ou télécharger une copie locale.
Lisez SKILL.md et les fichiers associés affichés par SkillsMP avant de décider de l'installer.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Une commande directe contourne le prompt de vérification. Examinez la source avant de l'exécuter.
npx skills add https://github.com/jessedegans/breather --skill backLa commande reste sur une seule ligne. Faites défiler horizontalement pour la vérifier avant de la copier.
Vous préférez une copie locale ? Téléchargez les fichiers actuellement disponibles dans 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.
Basé sur la classification professionnelle SOC
Affichage de 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.