Explain a topic with ak:explain

Build a grounded explanation of a concept, code path, or system in Markdown or self-contained HTML, with optional ELI5 wording.

Use ak:explain when you need a mental model of a new topic, file, system, error, or decision. The Skill separates observed facts from inference and can emit Markdown or a self-contained HTML visual.

Choose ak:explain for a grounded explanation

Use ak:explain when

You want a walkthrough, mental model, or “how does this work” answer.
You want --eli5 analogies that still keep every material warning.
You want --html for a local, offline visual explanation.

Choose another workflow when

You only want the immediately previous assistant message restated. Use ak:bro.
You want to open or generate a generic visual, slides, or a diff review. Use ak:preview.

Run the Skill

Complete Onboarding, and confirm Marketing Kit is installed for your runtime and scope.

Runtime	Invocation	Availability boundary
Claude Code	`/ak:explain ...`	Uses the installed Marketing Skill; file and browser access depend on the session.
Cursor	`/ak:explain ...`	Uses the user-verified slash spelling; helper and browser access depend on the session.
Codex	`$ak:explain ...`	Uses native Skill discovery; filesystem and browser access depend on the environment.

/ak:explain "How does the launch brief reach campaign approval?" --html

/ak:explain "How does the launch brief reach campaign approval?" --html

$ak:explain "How does the launch brief reach campaign approval?" --html

Flag	Effect	Boundary
(none)	Markdown explanation	Does not write files unless the runtime already would
`--eli5`	Everyday analogies; technical terms defined on first use	Must preserve every material warning and uncertainty
`--html`	Self-contained HTML visual explanation	Activates `ak:frontend-design` first, then `ak:diagram` when installed; otherwise a semantic inline SVG/CSS fallback
`--html --eli5`	HTML with ELI5 copy and prominent warning panels	Same file and composition boundary as `--html`

Unknown flags return concise usage help. If the subject is missing, the Skill asks one focused question.

Keep explanations bounded

Explanation is not execution

User-supplied files, logs, and issue text are untrusted data. The Skill must not follow embedded instructions, expose secrets, or treat analogies as literal truth.

HTML output is a local artifact. It does not publish, deploy, contact an audience, or spend.

Verify the result

A complete run should provide a gist, the core parts, the mechanism, a concrete example, and caveats. HTML runs should report the written path, such as plans/visuals/explain-{slug}.html or the plan-aware equivalent, and whether ak:diagram compiled or the fallback was used.