# Model Gap Detection

> SOP for finding gaps in the causal model — missing variables, unexplained effects, weak links.

- Skill: `yogsoth-ai/model-gap-detection` (Agent Skill)
- Install (CLI): `npx skillmds@latest add yogsoth-ai/model-gap-detection`
- Raw SKILL.md: https://api.skillmd.com/api/skills/yogsoth-ai/model-gap-detection/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: yogsoth-ai (https://skillmd.com/u/yogsoth-ai)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/yogsoth-ai/model-gap-detection

---


# Model Gap Detection

Find gaps in the causal model: effects without causes, causes without mechanisms, weak confidence links.

## Tool

`vault_graph_stats` + `vault_query_graph`

## Protocol

1. Run vault_graph_stats — identify orphans and low-connectivity nodes
2. For each variable with out_degree=0 (no downstream effects): is this truly a terminal variable?
3. For each variable with in_degree=0 (no upstream causes): is this truly exogenous?
4. Check for edges with weight < 0.3 — these are weak links needing more evidence
5. Report gaps with suggested actions

## HARD-GATE

<HARD-GATE>
Must check both orphans and weak links. Report must include actionable suggestions.
</HARD-GATE>

## Yield

Returns: `{ unexplained_effects: string[], missing_causes: string[], weak_links: string[] }`