Skip to content

Agent skills

Agent Skills are task-specific, progressively disclosed instructions distributed with this repository. They complement the API docs by encoding source inspection, decision flow, invariants, security boundaries, verification, and handoff requirements that are easy for an agent to get wrong during implementation.

Healthcheck skills

Seven Skills cover manager design, custom checks, diagnostics security, cgroups, exporters, Next.js routes, and Payload integration.

Payload Fields skills

Four Skills cover semantic factories, managed slugs, Jalali dates, and minor-unit money contracts.

Payload Editor skills

Three Skills cover core semantic configuration, preset contracts, and native Payload extensions.

Payload Schema skills

Two Skills separate entity consumption and migration from package compiler, adapter, and release development.

Payload Operations skills

Three Skills cover operation contracts, the shared Payload SDK client, and secure custom endpoints.

Install a Skill when the task matches its trigger. For example, install healthcheck-nextjs-routes when adding a Next App Router health route, not the entire collection by default.

Every Skill is self-contained for its primary task. A sibling Skill may be recommended for composition, but the primary workflow does not depend on reading it. Committed behavior evals cover a happy path, edge case, security/failure case, diagnosis/review case, and a near-miss or composition boundary; trigger evals balance 10 positive and 10 neighboring negative queries.