Publication Pass

SkillMedia

Decide whether an authored piece is ready to represent its writer in public by testing recognition, legibility, integrity, and worth. Use before publishing a post, video, essay, or controversial idea; not for platform formatting or audience approval.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Publication Pass skill

What this skill tells your AI

The instructions your AI receives, as published by epicenterhq/epicenter in .agents/skills/publication-pass/SKILL.md and read by ahel’s review.

Purpose

Decide whether a piece is ready to share publicly. It should sound like the writer, communicate what they actually mean, hold up under fair scrutiny, and be worth the attention and exposure it may bring.

This pass does not try to make everyone agree or sand off the sharp edges. The question is whether people will disagree with the writer's real view, rather than with something the writing accidentally suggests.

Read the Whole Piece

Read the artifact in the form people will encounter it. For a video, consider the words, narration, visuals, title, cover, caption, and post copy. Keep the spoken claim separate from display-only material, and notice when packaging makes a stronger promise than the piece itself.

Four Tests

Recognition

Does the writer recognize this as their own view?

Check the force of the claim, emotional posture, examples, uncertainty, and ending. Do not make the writer sound more neutral, certain, universal, or respectable than they are. A strong opinion can pass; a borrowed posture cannot.

Legibility

Can a good-faith stranger understand what the writer means without filling in the gaps with a different claim?

Name the central point, what the piece is not claiming, and any ambiguity that would invite a predictable false reading. Add a qualification only when it makes the intended claim more precise—not just to preempt every objection.

Integrity

Do the claim, examples, evidence, confidence, and qualifications fit together?

Check for unsupported facts, overgeneralization, selective examples, confused categories, and conclusions the piece has not earned. Separate disagreement from error. A controversial claim can pass when its scope and confidence are honest.

Worth

Is this worth the attention, exposure, and possible disagreement required to publish it?

Look for a live perception, concrete stake, useful distinction, or earned personal truth. Do not demand novelty for its own sake. A familiar idea can be worth sharing when it is expressed with unusual honesty or clarity.

When the Piece Is Controversial

Show the disagreement clearly:

  • the strongest charitable reading;
  • the strongest reasonable objection;
  • the most likely bad-faith misreading;
  • whether each problem comes from the writing or from the reader's projection;
  • the smallest clarification that would repair a writing-caused misreading.

Do not quietly soften the piece. If the remaining disagreement is a legitimate consequence of the writer's position, say so and leave it intact.

Judgment

Return one outcome:

  • Publish: It sounds like the writer, stands on its own, and is worth the exposure.
  • Publish with qualification: The core is ready, but one precise clarification is needed.
  • Revise: The idea is worth sharing, but the current piece is not yet faithful, clear, or earned.
  • Do not publish: It is not worth the exposure right now, or it captures a passing reaction rather than a view the writer wants to make public.

Name the decisive reason, the strongest remaining risk, and the smallest next move. If a wording change is needed, show it and wait for the writer to adopt it before editing authored prose.

Review Output

Return a compact, actionable pass:

  1. State the judgment first.
  2. Give the public claim in one sentence.
  3. Say whether it is recognizably the writer's view.
  4. Give the strongest charitable reading and strongest reasonable objection.
  5. Name any serious misreading and whether the writing causes it.
  6. Identify the smallest needed qualification or revision, if any.
  7. Say what should remain sharp or untouched because softening it would betray the piece.

Do not confuse “people may disagree” with “the piece is unclear.” The goal is for criticism to land on what the writer actually chose to say.

Signals

GitHub stars
5k
Forks
378
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
publication-pass
Source
github.com/epicenterhq/epicenter