← ClaudeAtlas

baton-inventorylisted

Map the work already on this machine into Baton task folders. Use once right after installing Baton, or when asked to find where past work lives, to categorise it and to reconstruct the logbooks.
StanimirTenev/baton · ★ 0 · AI & Automation · score 72
Install: claude install-skill StanimirTenev/baton
# Baton inventory — map the work that already exists Baton starts empty, but the machine does not. This walks the machine once, finds the work that was done before Baton existed, and turns it into task folders — **after the human confirms**. Nothing is moved, and no folder is created before that confirmation. Task root and logbook name: the session-start board names the task root in its first line ("Baton — the tasks in …"); installed as a plugin, the rules that come with it name both. With the install script they are also in `~/.claude/baton/hooks/baton.local.json` (`home`, `logbook`; defaults `~/tasks`, `LOGBOOK.md`). ## 0. Start from the survey Run the read-only survey first — `python3 <repo>/tools/baton_survey.py`, or `python3 "${CLAUDE_PLUGIN_ROOT}/tools/baton_survey.py"` in plugin mode. It names every memory folder, the places memory and instructions point at (other drives included), dated logbook files and folders that already keep a logbook. Hand those to the agents below: agent 3 reads **every** memory folder it lists, agent 4 also covers its places outside the home directory, and the dated logbook files are the record to reconstruct from, grouped by their topic prefix. ## 1. Scan — four read-only agents in parallel Launch four read-only (Explore) agents in one message. Each returns one table row per work item: `name (kebab) | paths | last activity | evidence | proposed state | next step | confidence`. Tell every agent: skip `node_modules`, virtualenvs, `.git`