kg-maintainlisted

Knowledge graph maintenance. Tend the garden — regular, light care keeps it healthy. Not a separate task: woven into every session, every capture. GARDEN RHYTHM — three modes, applied as needed: Water (routine): after each task, glance at 2–3 recently-touched nodes. Are their gists still accurate? Any note worth adding? Prune (when dense): merge duplicate nodes, shorten verbose gists (→ notes), split oversized nodes, remove stale touches, delete edges to removed concepts. Fertilize (on use): when a node proves valuable, connect it to newly-discovered related nodes. One new edge makes a node far more durable. AFTER CAPTURE: when you save a node, immediately ask — - Do any adjacent nodes now need updating? - Is this a duplicate of something existing? Merge if so. - Does this node's gist still fit, or did context shift? REACTIVE TRIGGERS (act immediately, mid-conversation): Uncertainty (spinning wheels, deja vu, about to search) → kg_search first. About to assume something → check KG; if missing, capture the ass
mironmax/claudecode-plugins · ★ 1 · AI & Automation · score 60

Install: claude install-skill mironmax/claudecode-plugins

# Maintenance Reference (Detailed) ## When Invoked Directly (/kg-maintain) Run a focused maintenance pass in this order: 1. `kg_read(cwd)` — check health stats: orphan %, avg edges/node, size warning 2. **Always**: scan all gists against the current kg-capture standard — tighten any that exceed it, regardless of graph size 3. **Always**: spot-check notes on recently-touched nodes — rewrite any that have grown chaotic or redundant (see "Notes Hygiene" below) 4. If graph is large or has size warning → **Prune**: merge duplicates, split oversized nodes, remove stale touches 5. After pruning → **Fertilize**: connect nodes clarified during pruning, add missing edges 6. **Water** throughout: update any gist that feels stale given what you just read Announce findings: "Graph health: N nodes, E edges, O% orphans. Running [prune|fertilize|water] pass." Report what changed: nodes merged, gists tightened, edges added. ## What a Healthy Graph Looks Like A healthy graph is a mesh of connections, not isolated facts. Most nodes participate in at least one edge. Health stats show this at a glance: - Low orphan rate — most nodes connected - Reasonable edge density — linked but not over-connected - Mix of levels — user patterns inform project decisions ## Maintenance Operations When auditing the graph with kg_read: - **Disconnected nodes** — appear in no edges. Connect them with edges if appropriate. Only delete if truly orphaned AND factually incorrect. - **Duplicates** — overlapping g