evolvelisted

<objective> Cluster active and proven instincts by lexical similarity and trigger family, write CLU-*.md cluster files, identify project instincts eligible for promotion, and update the INSTINCTS.md index with a Clusters section. </objective> ## Instructions You are clustering **existing** instincts — not extracting new ones and not promoting anything automatically. This is the structural pass that groups related behavioral rules so future skills can act on them. ### Step 1: Load Instincts Determine scope from the invocation arguments: - `--global` → only load from `~/.claude/instincts/global/` - `--project <slug>` → load from global AND `~/.claude/instincts/projects/<slug>/` - No flags → load global only (default to `--global` behavior) ```bash GLOBAL_DIR="$HOME/.claude/instincts/global" ls "$GLOBAL_DIR"/INS-*.md 2>/dev/null ``` If `--project <slug>` was passed, also load: ```bash PROJECT_DIR="$HOME/.claude/instincts/projects/<slug>" ls "$PROJECT_DIR"/INS-*.md 2>/dev/null ``` For each instinct file found, read the full content and parse the YAML frontmatter. Extract: - `id`, `name`, `status`, `confidence` - `signals.trigger_patterns` (list of strings) - `lineage.cluster_id` (current cluster assignment if any) - `score.successes`, `score.confirmations`, `score.opportunities` (or len of evidence array as opportunity proxy) - `evidence` (array length = opportunity count) - `created_at` - The `## Operational Rule` body text (the rule text used for similarity) **Filter:**