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

outline.mdskills/talk-oza-automating-design-quality/references/

Source outline

Source: Hamza Oza, Automating Design Quality in the Dark Factory, AI Native DevCon New York 2026. Recording. All anchors refer to original numbered cues in the transcript. Short quotations below preserve the supplied transcription, including spoken wording.

TopicSource anchorExact short quoteApplication and scope
Quality on the line00:04:30–00:06:01, cues 64–85“build the quality into the line”Earlier checks aim to reduce downstream rework.
Values voice visuals00:06:31–00:07:53, cues 94–118“these three pillars”Brand judgment includes more than visual tokens.
Composition gap00:07:56–00:09:11, cues 119–139“what do you leave out?”Component compliance does not establish good composition.
Objective comparison00:09:17–00:11:07, cues 140–169“an objective list of what are the key differences”Compare expert revisions before extracting recurring rules.
Advisory gates00:11:54–00:13:15, cues 181–200“These were designed to be advisory checks”The design checks were intentionally nonblocking.
Feedback to CI00:13:19–00:13:53, cues 201–209“mining for these rules”Repeated review feedback can become reusable guidance and checks.
Maintenance and reuse00:13:58–00:15:36, cues 210–235“creating these maintenance loops.”The workflow also addresses historical drift and non-code artifacts.
Coverage and exceptions00:15:38–00:17:11, cues 236–262“good design breaks the rules in appropriate places.”Rule coverage remains evolving and exceptions matter.
Prototype qualification00:18:59–00:20:12, cues 288–307“Sometimes we choose not to because you don't want the fidelity straight away,”Prototyping may deliberately use different fidelity.
Exception Q&A00:20:18–00:22:36, cues 308–345“we don't enforce all the checks”Exceptions inform discussion and contextual refinement.
Activation and evidence00:26:05–00:27:37, cues 393–418“it wasn't being picked up as often as we'd like”Skill availability does not guarantee activation.
Starting exercise00:27:44–00:28:39, cues 419–432“Pick 2 or 3 things”Start with a few recurring failures and expert corrections.

Source limitations

  • The examples, 60% shepherding-cost observation, and first maintenance findings concern Tessl's reported experience; they are not controlled savings estimates.
  • The transcript does not include the actual screen pixels, design tokens, full skill, or runnable CI configuration.
  • Tool capabilities and integration difficulties in Q&A are time-bound experience, not current documentation.
  • Rules are contextual and checks advisory in the described workflow; the source does not justify universal aesthetic rules or automatic release blocking.
  • Promotional interlude cues 170–180 are omitted from the bundled transcript; the remaining cue numbers retain their source positions.
  • This is a spoken transcript; unseen slides, linked repositories, and third-party research were not independently inspected. No outcome beyond the described talk is established.

Reference omissions

The agent-facing transcript omits unrelated promotional offer/sign-up cues 279–283. Original cue numbers and timestamps for substantive content are unchanged.

skills

talk-oza-automating-design-quality

tile.json