From trantor
Finish the CURRENT session in one move: write a rich model-authored handoff, open a fresh full-window session that takes over (it auto-recaps the handoff), and close this one once the fresh session has it — a clean baton pass, one session at a time. Use when you want to wrap up and continue fresh on demand (not just at the compaction threshold). Trigger: /trantor:handoff
How this skill is triggered — by the user, by Claude, or both
Slash command
/trantor:handoffThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Write a complete handoff capturing everything a NEW session needs to continue this work without
Write a complete handoff capturing everything a NEW session needs to continue this work without re-deriving context, and save it so the next session in this project auto-loads it on start.
Compose a thorough markdown handoff for the current task with these sections (be specific — exact file paths, concrete next actions; the successor has a fresh window and only this):
Save it AND pass the baton in one shot — pipe the markdown to the helper with --baton:
cat << 'HANDOFF' | node "${CLAUDE_PLUGIN_ROOT}/bin/write-handoff.mjs" --baton
<your handoff markdown>
HANDOFF
--baton writes the handoff, opens a FRESH session that takes over (it auto-recaps the handoff
on open), and closes THIS Terminal window once the fresh session has consumed it — a true baton
pass, one session at a time. (Omit --baton to only write the handoff without spawning/closing.)
It's safe: the original window is closed ONLY after the fresh session confirms it took over, and
never if the fresh session fails to start.
Tell the user briefly: "Handoff written — a fresh session is opening and will recap it; this window closes once it takes over." Then stop (the baton will close this session shortly).
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub sashabogi/trantor --plugin trantor