faviovazquez/sprang/sprang-health

Full structural health report for the knowledge graph.

Prerequisite: .sprang/knowledge-graph.json must exist. If it does not, run the sprang-analyze skill first.

Instructions

1. Call sprang_health to retrieve the full health report.
2. Lead with the health grade: display health_grade (A–F) and

health_score (0–100) prominently. Show the grade_breakdown — which factors cost the most points (dead code, circular deps, god nodes, coupling, security).

1. Display graph phase (skeleton or complete), generated_at,

total_nodes, and total_edges.

1. Present the risk distribution table from risk_summary: high, medium, low.

Note if high-risk nodes exceed 10% of total.

1. List the top 10 risky nodes (top_10_risky_nodes) with label, type, risk

score, and risk factors. Call sprang_node on any with god_node or circular_dependency warnings for detail.

1. List all detected code smell categories from smell_summary with counts and

a one-sentence description of each smell type. Valid categories: god_node, circular_dependency, orphan_node, unclear_coupling, duplicate_logic, low_cohesion, unstable_interface, over_connected, name_duplicate, layer_violation.

1. Report security findings from security_summary: total count, breakdown

by severity (high/medium/low) and by category (hardcoded_secret, sql_injection, xss_risk, etc.). Flag any high severity findings as immediate priorities.

1. Report orphan_count (isolated nodes), circular_dependency_count, and

nodes_without_tests as actionable gaps.

1. If history has entries, show the score trend (first vs latest snapshot) to

indicate whether health is improving or degrading.

1. Recommend a prioritized action list: high-severity security issues first,

then smells to address, then nodes needing tests, then when to re-run the sprang-analyze skill.