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-lee-perkins-code-readiness/references/

Codebase readiness and workflow slice

This worksheet is a recommended application of the joint talk, not a runnable workflow.

Repeated intervention

Describe one step a developer repeatedly performs or corrects. State why it is suitable for automation and which judgment remains with people.

Current evidence: Include the actual setup failure, repeated prompt, skipped task or review burden.

Runnable environment

Record the clean-start path, dependencies, seeded roles and required test data. Mark what was actually exercised.

List local substitutes for outside services and their limitations. Separate missing test setup from permission to access production.

Step contract

StepScript or agentRequired input/outputDefinition of doneFailure behavior
Fill with the selected sliceExplain whyConcrete artifactObservable checkBounded retry or escalation

Where a deterministic converter or check is required, missing execution must fail visibly. The agent must not silently improvise and report that the required tool ran.

Standards and verification

Identify useful architecture/context guidance and the standards enforceable by linting, types, formatting or tests.

Use mechanical checks for mechanical properties and a separate judge for plan compliance where useful. State the judge's limits and the human decisions still required.

Choose a retry budget appropriate to this task. Preserve required CI checks and record actual outcomes separately from the agent's summary.

Review evidence

Map each requirement to its status and supporting evidence. Use screenshots or short video only where they help assess changed behavior. Include meaningful check results and unresolved gaps.

Adoption decision

Measure intervention reduction together with regressions, rework and post-merge failures. State the acceptance condition, owner and next review point.

End with keep, revise or stop this one workflow slice. Do not generalize a single success into codebase-wide readiness or claim planned checks have passed.

skills

talk-lee-perkins-code-readiness

tile.json