← ClaudeAtlas

retirelisted

Retire a heavy executor: close it AND write a successor-seed.md indexing its packets/reports, so spawning a fresh session over the same territory is cheap. Invoke with /relay:retire, or when asked "retire that session", "this exec is getting heavy", "rotate this executor", "close it but keep what it learned".
spacegrowth/claude-relay · ★ 0 · Data & Documents · score 75
Install: claude install-skill spacegrowth/claude-relay
Call relay as `${CLAUDE_PLUGIN_ROOT}/bin/relay` (Claude Code substitutes the plugin's absolute path when this skill loads) — not bare `relay`, which often isn't on the Bash tool's non-interactive PATH. Run: `${CLAUDE_PLUGIN_ROOT}/bin/relay retire $session_id` That closes the session (as `superseded` by its seed, tab and all — same teardown as `/relay:close`) and writes a **successor seed** to its state dir: an index of every packet it was sent, with each report's outcome line, `Status`, `Risk flags` and `UNVERIFIED` — plus the worktree and topic it owned. The seed is an INDEX, not a transcript: the detail stays in the linked reports. **Then spawn the successor with the seed**, exactly as the retire message tells you: `${CLAUDE_PLUGIN_ROOT}/bin/relay spawn $worktree $topic $packet --seed $retired_session_id --lead "$CLAUDE_CODE_SESSION_ID"` The seed is appended to the fresh executor's packet as inherited context (task first, seed second, GATES last), so the successor starts knowing what happened on this territory without reading anything else. You can also pass a path to the seed file instead of the retired session's id. **When to reach for this instead of `/relay:send`.** A long-lived executor accumulates context; a fresh one starts sharp. Retire when a session is getting heavy, has finished a phase of work, or is dead/stalled with useful history behind it — the seed is what makes respawning cheap enough to actually do, so don't keep piling packets onto a heavy session