CtrlK
BlogDocsLog inGet started
Tessl Logo
PluginPublic

ainativedev/aidevcon-2026-nyc

AI Native DevCon 2026 New York — practical agent skills grounded in conference talks

Quality

86%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Overview
Quality
Evals
Security
Files

application-template.mdskills/talk-oza-automating-design-quality/references/

Design judgment extraction record

This worksheet applies Oza's comparison method. Use actual artifacts and preserve the distinction between observation and proposed rule.

Comparison set

Task and intended audience: Generated artifact: Expert revision: Other paired examples: Source of brand values and voice: Fidelity appropriate to this stage:

Use a few recurring failures. A rough prototype may be appropriate if the current question concerns the product idea.

Objective differences

PairObservable changeUser-facing effectRepeated elsewhere?

Describe hierarchy, grouping, content removal, density, spacing, and emphasis where visible. Do not begin with a generic “better design” score.

Candidate rule

Rule: Evidence across pairs: Applicable context: Reason it helps: Known exceptions: Confidence: New example for validation:

Separate token compliance from composition. A rule extracted from a launch page should not automatically apply to every product screen.

Delivery and verification

Before-generation guidance: How activation is detected: Deterministic checks: Visual or judgment checks: Advisory finding format: Before/after evidence: Maintenance opportunity:

Do not claim a check is deterministic when it depends on subjective interpretation. Keep advisory preferences separate from requirements already mandated by the project.

Exception learning

Flag raised: Reason the output differs: Intentional exception or poor rule: Who supplied the contextual judgment: Narrow scope of exception: Change to reusable guidance: Regression or comparison case:

Preserve the original rule where it still applies. Avoid broadening permission from one exception.

Review result

Observed improvement: Repeated feedback reduced: New problem introduced: Unknowns: Smallest next revision:

Mark checks and comparisons as proposed until they are actually performed. This artifact supports learning from design work; it does not automatically schedule maintenance, modify CI, or publish a design.

skills

talk-oza-automating-design-quality

tile.json