Slop Doctrine Refresh And Research Loop

Slop Doctrine Refresh & Research Loop

Acknowledgement: Shared by Peter Bamuhigire, techguypeter.com.

Use When

• A user says the definition of "AI slop" has changed.
• A font, visual style, layout trope, AI-image tell, or UI/product pattern may have become a new AI-default signal.
• Updating ai-slop-taxonomy.md, ai-slop-banned-fonts.md, visual-product-slop-audit, or ai-slop-typography-audit.
• A high-stakes public artifact needs a freshness check against current AI-slop norms.
• Another engine asks whether a design choice is merely popular, genuinely AI-converged, or still acceptable with a deliberate rationale.

Do Not Use When

• You are auditing a specific image/UI only against the current checklist - use visual-product-slop-audit.
• You are only auditing type in a finished artifact - use ai-slop-typography-audit.
• The concern is written copy only - route to the digital-research engine's anti-ai-slop and ai-slop-audit skills.

Required Inputs

• The doctrine, checklist, or artifact under review.
• The suspected changed tell, default, font, product trope, or market signal.
• The intended scope: typography, image/video, UI/product, brand, document, or cross-engine handoff.
• Whether the result is advisory, watchlist-level, or intended to change a hard ban.

Workflow

1. Load the local anti-slop doctrine: doctrine/design-doctrine.md, doctrine/references/ai-slop-taxonomy.md, doctrine/references/ai-slop-banned-fonts.md, and doctrine/references/living-slop-refresh-protocol.md.

2. Route research to the digital-research engine before making doctrine changes. Load its skills/source-evaluation/SKILL.md, skills/source-evaluation/references/evidence-discipline.md, skills/anti-ai-slop/SKILL.md, and skills/ai-slop-audit/SKILL.md.

3. Separate claims by evidence type: primary/vendor/platform source, design-practitioner source, public incident, marketplace/template observation, social signal, or weak anecdote.

4. Classify the change using the protocol buckets: hard ban / critical tell, watchlist, contextual risk, or rejected claim.

5. Translate only the design consequence into this engine. Keep writing/cultural taxonomy in the digital-research engine and cross-cite it instead of duplicating it.

6. Patch doctrine with a dated rationale if the evidence warrants it. State observed shift, evidence grade, design consequence, scope, and date checked.

7. Run the design quality gate and verify every active SKILL.md still has an example when a new skill is added.

8. Apply the machine-error overlay: review whether a new visual or product pattern adds a user task, state, decision, proof point, or meaningful hierarchy. Record ME1-ME7 findings in the proposal delta; route prose semantics to Digital Research and mark incomplete evidence NOT_ASSESSED.

9. Enforce the visual boundary: for websites, web products, presentations, and design artefacts, reject purple gradients, glassmorphism, neon glow, AI-beige defaults, decorative editorial scaffolding, and decorative motion. Preserve only task, state, accessibility, data, or approved design-system reasons, and keep unavailable evidence NOT_ASSESSED.

Anti-Patterns

• Freezing a 2026 slop tell as permanent truth after models have moved.
• Adding hard bans from vibes, social screenshots, or a single anecdote.
• Treating AI-vendor design recommendations as approval authority.
• Copying the digital-research engine's writing guidance into this repo instead of referencing it.
• Updating a checklist without recording evidence strength and date checked.

Outputs

• A slop-doctrine refresh note: observed shift, evidence grade, design consequence, scope, date checked, and changed files.
• Updated design doctrine/checklists when evidence supports a change.
• A rejected-claim note when the evidence is too weak for doctrine.

Examples

• Use the worked example in examples/ for the source register and doctrine-delta format.

Quality Standards

• Use current, traceable evidence and distinguish human design authority from AI-vendor self-report.
• Require corroboration before adding a ban or universal tell.
• Stop publication when evidence conflicts materially; retain the current doctrine and record uncertainty.

Decision Rules

Capability Contract

Read, search, and network research are required; doctrine editing requires explicit authority. Preserve citations and change history. Do not update bans from model memory, social virality, or a single source.

Degraded Mode

If required evidence or tooling is unavailable, use the scoped fallback below and mark the result unverified. Without current research access, return the question set and source plan, keep doctrine unchanged, and label all currency claims unverified. Recover when authoritative sources can be checked.

• examples/font-default-refresh-note.md - classifies an emerging AI-default font signal and decides whether it becomes a hard ban, watchlist item, or rejected claim.

References

• doctrine/references/living-slop-refresh-protocol.md - required update protocol.
• doctrine/references/ai-slop-taxonomy.md and doctrine/references/ai-slop-banned-fonts.md - design-engine slop doctrine.
• visual-product-slop-audit and ai-slop-typography-audit - artifact-level audits that consume the refreshed doctrine.
• Digital-research engine: source-evaluation, evidence-discipline, anti-ai-slop, and ai-slop-audit - source discipline and writing/cultural slop framing.