source-credibility-analyzer

---

LIBRARY-FIRST PROTOCOL (MANDATORY)

Before writing ANY code, you MUST check:

Step 1: Library Catalog

• Location: .claude/library/catalog.json
• If match >70%: REUSE or ADAPT

Step 2: Patterns Guide

• Location: .claude/docs/inventories/LIBRARY-PATTERNS-GUIDE.md
• If pattern exists: FOLLOW documented approach

Step 3: Existing Projects

• Location: D:\Projects\*
• If found: EXTRACT and adapt

Decision Matrix

Match	Action
Library >90%	REUSE directly
Library 70-90%	ADAPT minimally
Pattern exists	FOLLOW pattern
In project	EXTRACT
No match	BUILD (add to library after)

---

STANDARD OPERATING PROCEDURE

Purpose

• Evaluate sources before using them in research outputs.
• Make constraints and evidence explicit to avoid unvetted citations.
• Keep structure-first documentation aligned with skill-forge standards.

Trigger Conditions

• Positive: screening papers/articles/web data for reliability; comparing conflicting sources.
• Negative: general synthesis (use general-research-workflow) or idea generation without sourcing.

Guardrails

• Constraint buckets: HARD (recency, peer-review, provenance), SOFT (domain experts, diversity), INFERRED (author incentives) with sources.
• Two-pass evaluation: factual/provenance checks → epistemic/impact assessment.
• Confidence ceilings: observation 0.95 for verifiable facts; inference 0.70 for bias assessments.

Inputs

• Source metadata (authors, venue, date, links).
• Research context and required quality thresholds.

Workflow

1. Scope & Constraints: Record quality criteria and constraint buckets; confirm INFERRED motivations/risks.
2. Provenance Check: Validate authorship, venue, recency, and originality.
3. Content Check: Inspect methodology, data quality, reproducibility, and alignment to question.
4. Bias & Risk Assessment: Identify conflicts of interest, cherry-picking, or missing context.
5. Score & Recommend: Rate credibility with rationale and ceiling; recommend use/avoid/follow-up; store notes in references/examples.

Validation & Quality Gates

• Evidence recorded for scores; conflicting indicators highlighted.
• Constraints respected; red flags noted.
• Confidence ceilings attached to each rating.

Response Template

**Source & Context**
- ...

**Assessment**
- Provenance: ...
- Methodology: ...
- Bias/Risks: ...

**Rating**
- Credibility score + rationale + confidence ceiling.
- Recommendation: use / cautious / avoid.

Confidence: 0.82 (ceiling: inference 0.70) - based on evaluated criteria and evidence.

Confidence: 0.82 (ceiling: inference 0.70) - reflects completed checks with documented evidence.