Story Arborist

Story Arborist - Tree Structure Analysis and Reorganization

Diagnose structural and semantic issues in story-tree.db using a two-phase approach:

1. Phase 1: Structural Health — deterministic scripts check closure table integrity
2. Phase 2: Semantic Organization — AI analysis of node hierarchy and grouping

Database: .claude/data/story-tree.db Scripts: .claude/skills/story-arborist/scripts/

Design Philosophy

Analysis is hard; execution is mechanical. This skill:

1. Focuses Claude's context on analyzing tree structure and identifying issues
2. Delegates all mechanical operations (moves, renames, path rebuilds) to Python scripts
3. Scripts run without loading into context, preserving budget for diagnosis

Quick Reference

All scripts support --dry-run for preview.

Workflow: Diagnose Then Fix

Step 1: Run Health Check

python .claude/skills/story-arborist/scripts/tree_health.py

Reports: statistics, stage/hold/terminus distribution, structural issues.

Step 2: Analyze Issues

For each issue category, determine root cause:

Step 3: Apply Fixes

Always use --dry-run first:

# Preview fix
python .claude/skills/story-arborist/scripts/move_node.py 8.6 root --dry-run

# Apply fix
python .claude/skills/story-arborist/scripts/move_node.py 8.6 root

Step 4: Verify

python .claude/skills/story-arborist/scripts/validate_tree.py

Phase 2: Semantic Organization Review

After structural issues are resolved, analyze semantic organization. This phase requires AI judgment—scripts can only surface data; you must interpret it.

Semantic Health Indicators

Diagnostic Queries

Run these queries against .claude/data/story-tree.db to surface semantic issues:

Leaf Epics (depth-1 with no children)

SELECT n.id, n.title, n.stage, n.terminus
FROM story_nodes n
JOIN story_paths p ON p.descendant_id = n.id AND p.ancestor_id = 'root' AND p.depth = 1
WHERE NOT EXISTS (
    SELECT 1 FROM story_paths p2 WHERE p2.ancestor_id = n.id AND p2.depth = 1
);

Nodes returned here are primary epics with no children—likely implementation details masquerading as epics.

Depth-1 Distribution

SELECT n.id, n.title, n.stage, n.terminus,
    (SELECT COUNT(*) FROM story_paths p2 WHERE p2.ancestor_id = n.id AND p2.depth > 0) as descendants
FROM story_nodes n
JOIN story_paths p ON p.descendant_id = n.id AND p.ancestor_id = 'root' AND p.depth = 1
ORDER BY descendants;

Look for:

• 0 descendants: Leaf epics—should be demoted
• Uneven distribution: Some epics may be absorbing too much or too little

Rejected/Archived at Depth-1

SELECT n.id, n.title, n.terminus, n.stage
FROM story_nodes n
JOIN story_paths p ON p.descendant_id = n.id AND p.ancestor_id = 'root' AND p.depth = 1
WHERE n.terminus IN ('rejected', 'archived') OR n.hold_reason IS NOT NULL;

These nodes clutter the primary epic list and should be moved to an archive area.

Epic vs Feature vs Task Guidelines

Red flags:

• Epic with 0-1 children → probably a feature misclassified
• Depth-1 title with implementation specifics → wrong granularity
• Multiple depth-1 nodes with overlapping keywords → fragmentation

Semantic Analysis Workflow

1. Run diagnostic queries to surface potential issues
2. Group by theme — identify related nodes that might consolidate
3. Propose reorganization — draft moves/merges with rationale
4. Validate with user — semantic changes are subjective; get approval
5. Execute with scripts — use move_node.py, bulk_reparent.py, etc.

Common Scenarios

Misplaced Root Children

Symptom: Decimal IDs like 8.6, 8.7 appearing as direct children of root.

Diagnosis: These should be integer IDs (16, 17, etc.) since root children must be primary epics.

Fix options:

1. Move to root (auto-assigns integer ID):

   python scripts/move_node.py 8.6 root
   

2. Rename directly:

   python scripts/rename_node.py 8.6 16
   

Orphaned Nodes

Symptom: Stories exist in story_nodes but are invisible to tree traversal.

Diagnosis: Missing story_paths entries.

Fix:

python scripts/fix_orphans.py

Bulk Reorganization

Symptom: Multiple nodes need to move to same parent.

Fix:

python scripts/bulk_reparent.py 15 1.2 1.3 1.4

Corrupted Paths

Symptom: parent_mismatch or missing_self_paths errors.

Fix:

python scripts/rebuild_paths.py root --recursive

ID Format Rules

Critical: Primary epics (root children) MUST have plain integer IDs, never decimal.

Script Details

tree_health.py

Full diagnostic report with statistics and structural issues.

• --json: Output machine-readable JSON

validate_tree.py

Structural validation only, no statistics.

• --json: Output as JSON

list_orphans.py

List orphaned nodes.

• --ids-only: Output IDs only (for scripting)

fix_orphans.py

Auto-fix all orphaned nodes by rebuilding paths.

• --dry-run: Preview without changes

create_node.py

Create new node under a parent with auto-generated ID.

• --title: Required. Node title
• --description: Optional. Node description
• --stage: Stage (default: concept)
• --capacity: Story capacity/points
• --dry-run: Preview without changes

move_node.py

Move node to new parent with auto-generated ID.

• --dry-run: Preview the move

rename_node.py

Rename node and all descendants.

• --dry-run: Preview renames

rebuild_paths.py

Reconstruct closure table entries.

• --recursive: Include descendants
• --dry-run: Preview

bulk_reparent.py

Move multiple nodes to same parent.

• --dry-run: Preview moves

Database Utilities

Scripts use shared functions from story-tree/utility/story_db_common.py:

• validate_tree_structure(conn) - Find issues
• rename_story(conn, old_id, new_id) - Rename with descendants
• rebuild_paths(conn, node_id) - Single node
• rebuild_paths_recursive(conn, node_id) - Node and descendants
• move_story(conn, story_id, new_parent_id) - Move with validation
• get_next_child_id(conn, parent_id) - Next available child ID