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-cunniffe-reusable-agent-platform/references/

Reusable platform slice worksheet

This is a recommended application of Cunniffe's argument, not a schema supplied in the talk.

Decision and current work

Repeated task: Describe one real request, its users and intended result.

Human bottleneck: Record what consumes review or maintenance time. Distinguish environment setup from repeated app logic. Include a concrete example and mark missing evidence.

Candidate primitive: Name one reusable capability. Explain why its boundary is coherent and what remains feature-specific. State a condition under which building it would cost more than it saves.

Interface and feature composition

Contract: List inputs, outputs, failure behavior and data access. Identify what callers may customize and what they must obtain through the interface.

Composition: Show the feature as a short chain of existing and proposed primitives. For a dashboard, consider query behavior separately from chart behavior and layout. Use descriptive placeholders until actual APIs are inspected.

Agent guidance: Give the interface location, one normal usage example and how to request a contract extension. Distinguish reuse from permission to alter the component.

Verification and contribution boundary

Primitive checks: Name meaningful failures and evidence that shared behavior works.

Feature checks: State what still needs verification after composition, including access and integration behavior.

Contributor scope: Describe permitted changes, sensitive boundaries and the existing review path for exceptions. For an internal app, trace both the app's permissions and those inherited by people accessing it.

Bounded experiment

Choose one representative task and compare duplicate logic, review effort, unexpected component changes and behavior verified. Define acceptance criteria suitable for the team; mark them proposed rather than measured.

Decision: Adopt, revise or abandon this slice, with evidence and unresolved questions. Do not claim platform-wide savings from one successful example.

skills

talk-cunniffe-reusable-agent-platform

tile.json