Give your show one look
Define who should watch and why this show is worth their time. Turn that choice into rules the team can use each week. Leave a shared profile so later work can start from known facts.
Start from the show
Read SHOW-PROFILE.md in the show’s working folder if it exists. It’s a short notes file about the show (audience, look, past titles, and lessons) that the Podcasting skills share. Keep confirmed choices unless the user asks to change them. Update only what this task resolves.
| Input | Action |
|---|---|
| Channel link only | Inspect the channel, its description, and a sample of recent episodes. Aim for 12 to 20 title and thumbnail pairs. If images are unavailable, review words and ask for a contact sheet before judging visual detail. |
| Transcript only | Infer a possible viewer need and tone from the talk. Mark both as provisional. Ask whether this episode reflects the whole show. Leave visual rules as proposals. |
| No show yet | Ask who the show helps and what those people should get from it. Draft a starting direction from the answer. Do not invent an audience history or tested style. |
Use a smaller sample when that is all the user has. State the limit. Do not turn setup into a long form to fill out.
Choose the reason to watch
Write one sentence: “For [specific viewer], this show helps with [need] through [kind of conversation].”
Name the viewer’s situation in plain words. “People who want to grow” is too broad. “New managers who need help with hard talks at work” gives the team a choice it can use.
Find proof in the show: the host’s questions, guest access, lived experience, or how the subject is explored. Separate an ambition from a trait the show already has.
Find a wider human need without losing the core topic. A craft show can help people think clearly. The episodes must support that wider frame. Set a boundary so a famous guest or a hot topic cannot pull the show anywhere.
Choose how to stand out
Look at three to five shows the same viewer might choose. Inspect the actual titles and images. If the Benchmark MCP is connected, use it to find public posts, comparable shows, and outlier episodes. Otherwise use public pages or the user’s examples.
Record the dominant look and promise. Find a useful contrast this show can sustain. A measured conversation may stand out in a loud category. A playful show may stand out in a formal one. Do not force a quiet style onto every genre.
For each borrowed idea, say what it solves and how to express it in this show’s own style. Avoid lifting another show’s complete layout.
Set rules a designer can follow
Choose values that fit the real footage. For each rule, mark it as observed, confirmed, or proposed.
| Area | Decide |
|---|---|
| Images | Real shoot frames, portraits, or artwork; how to show the interview format |
| Faces | Who appears, rough share of frame, acceptable expressions, and when the host joins |
| Color | Base colors, accent use, and a repeatable photo treatment |
| Type | Font and weight, text zone, word range, and when to use no text |
| Layout | Crop, margins, text placement, and room for interface labels |
| Titles | Where names and credentials go, useful patterns, and words the show avoids |
| Format cues | How to tell full episodes from clips while keeping a shared look |
Match the visual weight to the content. Use real detail, good framing, and room to read when the show asks for close attention. Use energy when the conversation earns it.
Describe one base layout and a useful exception. Test both against several guests, including one without a known face or strong photo. Rules that work only for the star guest will fail next week.
Show how six recent or proposed pairs would sit together. Use a contact sheet if image tools are available. Otherwise give a labeled layout plan. State when visual checks remain open. Inspect both the row and the individual pair at phone size.
Choose catalog fixes
Review current pairs against the proposed rules. Keep strong work. Do not repaint the catalog just to make every image identical.
Prioritize episodes with lasting value, useful guest expertise, and a clear package flaw. Check views at a similar age when possible. A low view count alone does not prove the thumbnail failed. If private data is available, check whether people who click keep watching.
Give up to five fixes. Name the episode, current issue, exact change, and how to judge the result. Start with a small batch. Preserve current asset IDs so the team can compare or revert.
Save the shared profile
Load references/show-profile.md when creating or updating the profile. If you can’t open it, continue with the steps here. Fill it from evidence already gathered. Keep unknowns visible.
Use one SHOW-PROFILE.md per show, in the user’s working folder. For several shows, keep one folder per show. If file access is unavailable, return the profile as text the user can save. Do not create a second profile because another tool uses a different folder.
Return the decisions
Lead with the proposed audience and reason to watch. Then provide:
- How the show should stand out, with the evidence behind it.
- Style rules with values and a sample row of title and thumbnail pairs.
- A short fix list for the current catalog, if there is one.
- The completed show profile, including open choices.
For an invented example, use Better Questions, a calm interview show for busy adults. It explores the choices behind daily habits. Propose warm shoot photos, one main face, and up to three words in a fixed text zone. Its sleep-tracking episode with fictional researcher Dr. Lena Park should feel part of that show. Label these as proposed rules until the team adopts them.
If installed, use podcast-packaging next for a specific episode. Use podcast-clips if the immediate task is a release plan. This skill’s profile and rules remain usable on their own.