| name | dxpt - Domain Expert Assessment |
| description | Evaluate quality from a domain expert's perspective. Identifies the domain, checks for oversimplification, verifies terminology, assesses depth, identifies missing nuance, and flags misleading content. |
| output | {"format":"prose"} |
Domain Expert Assessment
Input: $ARGUMENTS
Step 1: Identify the Domain
Establish the relevant domain and the expert lens to apply.
SUBJECT: [what is being evaluated]
PRIMARY DOMAIN: [the main field of expertise needed]
ADJACENT DOMAINS: [related fields that inform evaluation]
EXPERT LEVEL NEEDED: [practitioner / specialist / researcher]
DOMAIN CHARACTERISTICS:
- Maturity: [established / evolving / nascent]
- Consensus level: [high agreement / active debate / fragmented]
- Key frameworks: [dominant models or theories in this domain]
- Common misconceptions: [what non-experts typically get wrong]
Step 2: Check for Oversimplification
Identify where complexity has been reduced to the point of inaccuracy.
OVERSIMPLIFICATION CHECK:
1. [claim/statement]
Simplified version: [what was said]
Full picture: [what an expert would say]
Severity: [HARMLESS / MISLEADING / WRONG]
Fix: [how to add nuance without losing accessibility]
2. [claim/statement]
Simplified version: [what was said]
Full picture: [full expert view]
Severity: [level]
Fix: [suggestion]
ACCEPTABLE SIMPLIFICATIONS:
- [simplification]: OK because [audience/context justifies it]
DANGEROUS SIMPLIFICATIONS:
- [simplification]: Leads to wrong conclusions because [reason]
Step 3: Verify Terminology
Check whether domain-specific terms are used correctly.
TERMINOLOGY AUDIT:
| Term Used | Correct Usage? | Standard Term | Notes |
|-----------|---------------|---------------|-------|
| [term 1] | [YES / NO / CLOSE] | [correct term] | [detail] |
| [term 2] | [YES / NO / CLOSE] | [correct term] | [detail] |
MISUSED TERMS:
- [term]: Used to mean [X], actually means [Y] in this domain
MISSING TERMS (should have been used):
- [concept] should be called [correct term] — matters because [reason]
INFORMAL VS FORMAL: [is the register appropriate for the audience?]
Step 4: Assess Depth of Treatment
Evaluate whether the topic is covered with sufficient depth.
DEPTH ASSESSMENT:
| Topic Area | Depth Provided | Depth Needed | Gap |
|-----------|---------------|-------------|-----|
| [area 1] | [surface / moderate / deep] | [what's appropriate] | [description] |
| [area 2] | [surface / moderate / deep] | [what's appropriate] | [description] |
UNDEREXPLORED AREAS:
1. [area] — Why it needs more depth: [reason]
Key points missing: [specifics]
OVEREXPLORED AREAS (unnecessarily detailed):
1. [area] — Why it's too deep: [reason]
Could be reduced to: [summary]
APPROPRIATE DEPTH OVERALL: [SHALLOW / ADEQUATE / THOROUGH / EXCESSIVE]
Step 5: Identify Missing Nuance
Find important caveats, exceptions, and context that experts would flag.
MISSING NUANCES:
1. [nuance] — Matters because: [impact on understanding]
Expert would add: [what they'd say]
2. [nuance] — Matters because: [impact on understanding]
Expert would add: [what they'd say]
MISSING CAVEATS:
- [statement] needs caveat: [what's true only under certain conditions]
MISSING CONTEXT:
- [claim] omits: [historical / cultural / technical context]
MISSING COUNTERARGUMENTS:
- [position] doesn't acknowledge: [legitimate opposing view]
Step 6: Flag Misleading Content
Identify anything that could lead readers to incorrect conclusions.
MISLEADING CONTENT:
1. [content] — Misleading because: [reason]
Reader would conclude: [wrong takeaway]
Correct framing: [how to present it accurately]
Severity: [CRITICAL / MODERATE / MINOR]
2. [content] — Misleading because: [reason]
Reader would conclude: [wrong takeaway]
Correct framing: [how to present it]
Severity: [level]
DOMAIN EXPERT VERDICT:
- Accuracy: [1-5]
- Depth: [1-5]
- Nuance: [1-5]
- Terminology: [1-5]
- OVERALL: [score] / 5
SUITABLE FOR AUDIENCE: [who can safely consume this as-is]
NEEDS REVISION FOR: [who would be misled by current state]
TOP PRIORITY FIXES:
1. [fix] — Addresses: [issue]
2. [fix] — Addresses: [issue]
3. [fix] — Addresses: [issue]
Integration
Use with:
/llmf -> Assess if domain expertise can be reliably provided by an LLM/nusr -> Balance expert accuracy against newcomer accessibility/upth -> Formalize the domain expert's implicit evaluation criteria
