# Strength Assessment

> Tactic for assessing argument strength — evaluate evidence quality, count independent sources, check for defeaters, assign calibrated confidence scores.

- Skill: `yogsoth-ai/strength-assessment-2` (Agent Skill)
- Install (CLI): `npx get-skill add yogsoth-ai/strength-assessment-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/yogsoth-ai/strength-assessment-2/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: yogsoth-ai (https://skillmd.com/u/yogsoth-ai)
- Updated: 2026-09-21
- Page: https://skillmd.com/skills/yogsoth-ai/strength-assessment-2

---


# Strength Assessment

Evaluate the strength of claims based on their evidential support.

## Minimum Yield

<HARD-GATE>
Must score ≥3 claims per invocation.
</HARD-GATE>

## Protocol

1. For each claim, gather all linked evidence (supporting and contradicting)
2. Evaluate evidence quality:
   - Source reliability (peer-reviewed > preprint > blog)
   - Directness (direct test > indirect inference > analogy)
   - Independence (multiple independent sources > single source chain)
3. Check for defeaters:
   - Undercutting defeaters (evidence that the support doesn't actually support)
   - Rebutting defeaters (evidence directly contradicting the claim)
   - Undermining defeaters (evidence that the evidence itself is unreliable)
4. Assign strength score (0-10) with explicit reasoning via `strength-scoring`
5. Flag claims with score < 3 as "weak" and claims with score > 7 as "strong"

## SOPs Used

- **strength-scoring** — assign and record strength scores
- **evidence-attachment** — create additional evidence links discovered during assessment

## Yield Report

Returns: `{ claims_scored: number, strong_claims: number, weak_claims: number, defeaters_found: number }`

