← ClaudeAtlas

meta-dream-crosstree-deeplisted

Use on "deep crosstree dream", "/dream-crosstree-deep", "deep cross-project scan", or when you want the exhaustive cross-project/cross-tree read regardless of whether anything obviously changed - the full semantic fan-out over ALL project memory stores AND their CLAUDE.md files, no convergence shortcut, no asking. For the normal, cheaper global dream that convergence-checks first and asks before the expensive scan, use meta-dream-crosstree.
bitranox/bitranox-skills · ★ 1 · AI & Automation · score 57
Install: claude install-skill bitranox/bitranox-skills
# meta-dream-crosstree-deep The exhaustive variant of `bitranox:meta-dream-crosstree`. Same goal, same safety model, same outputs - the ONE difference is that the **deep cross-project semantic scan is mandatory here, not opt-in**: this skill always reads every store (and every CLAUDE.md), even if the cheap convergence pre-check says nothing changed. Use it when you want the thorough read; use `bitranox:meta-dream-crosstree` for the routine, convergence-gated pass that asks before going deep. **REQUIRED BACKGROUND:** Follow `bitranox:meta-dream-crosstree` for the full procedure (capture-first reading the session from DISK via `dream_state.py session-review`, backup, inbound gather, promotion gate, outbound cross-pollination, re-dedup + reconcile, skill-fit, report) and `bitranox:meta-self-improve` for the altitude/normalization primitives. This skill replaces how the scan is run, and FOLDS IN crosstree's promotion gate and misplacement audit as its own steps 3 and 3b - so those two are already done when you reach step 5, and running them again is the one duplication to avoid. Everything else, do not duplicate: follow crosstree. ## What changes vs meta-dream-crosstree 1. **Back up first** (per-run snapshot of each affected tree's TOP store + any store you will write) - unchanged. 2. **Always run the semantic fan-out - no convergence shortcut, no asking.** FAN OUT one **`sonnet`** subagent per project store, OR (for many stores) one per thematic batch, in parallel. Each r