agentleFS
2104178150-ai2104178150-ai/megaphone/skills/launch-positioning/SKILL.md

launch-positioning

Skill5 starsChanged 32 days agoMITView on GitHub

Find the one-sentence pitch for an open-source project or app — who it's for, what it replaces, why now, the proof, how it beats today's alternatives, and what we must not claim — and save it as the launch brief every post is built from. Use when a repo's pitch is vague, before writing a README or launch posts, or when the user asks 怎么介绍我的项目 / 一句话说清楚 / 定位 / tagline / elevator pitch.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read .megaphone/decisions.md first if it exists, and record new answers there — or, with no human available, your assumptions (and mark …

    line 14
  2. Very specific:

    If the user isn't available, fill fields from the repo, mark them [inferred], and leave the origin story as [TODO] — never invent it.

    line 33
  3. Distinctive:

    Swap test: if you swap in a competitor's name, is it still true?

    line 63
  4. Distinctive:

    No-buzzword test: remove "AI-powered", "blazing", "next-gen", "revolutionary", 「神器」「颠覆」.

    line 68
  5. Very specific:

    Founder-voice test: every sentence in the author's voice ("I built it because…") is the author's own words or a [TODO].

    line 70
  6. Distinctive:

    Write it in the user's language.

    line 76
  7. Distinctive:

    Write the Chinese one-liner natively, not as a translation: Chinese audiences respond to concrete pain ("再也不用手动…") and a clear 「适合谁」.

    line 122

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Launch positioning

Most launch posts fail before they're written: the author can't say in one breath who the project is for and why it's better than what those people use today. This skill produces .megaphone/brief.md, the single source every other Megaphone skill reads.

Steps

Read .megaphone/decisions.md first if it exists, and record new answers there — or, with no human available, your assumptions (and mark inferred fields [inferred] in the brief).

  1. Read before asking. README (look for why, motivation, 初衷 sections), docs, examples, the oldest issues and the first commits often reveal the original itch. Note what the code actually does — features the README doesn't mention are common. Read .megaphone/audit.md if it exists and copy every do-not-say item from its risk section into Claims we won't make.
  2. Check the alternatives' current state. For each alternative you'll contrast with, look at its latest release and its open issues/PRs for the feature you're comparing on. A competitor may be about to ship it. Date-stamp every contrast claim ("as of 2026-10").
  3. Interview briefly (skip anything already answered):

    • What were you doing the day you started this? What was annoying?
    • Who else has that problem? Be specific: "Go developers who ship CLIs", not "developers".
    • What do they use today instead? (Another tool, a script, a spreadsheet, nothing.)
    • What's the moment it clicks — the first result that makes someone say "oh nice"?
    • What changed recently that makes this matter now?
    • Any proof: numbers you measured, users, a before/after?

    If the user isn't available, fill fields from the repo, mark them [inferred], and leave the origin story as [TODO] — never invent it.

  4. Fill the canvas below. Push for specifics; "fast", "simple" and "powerful" aren't positions — every README says them.
  5. Write three one-liners in different styles, recommend one, and say why:
    • Outcome: "Turn X into Y in one command."
    • Contrast: "Like A, but for B." (only if A is well known to the audience)
    • Pain: "Stop doing X by hand."
  6. Check them against the tests below, then save .megaphone/brief.md.

Canvas

Two example columns: a developer tool and an end-user app.

FieldQuestionDeveloper toolEnd-user app
AudienceWho exactly?maintainers of small open-source repos launching for the first time网易云 users with years of playlists they don't want to lose
TriggerWhen do they reach for it?the night before posting on Hacker Newstheir membership expires, or songs in a playlist go grey
Today they useWhat do they do now?skim old blog posts and guessscreenshots and retyping
PromiseWhat changes for them?a scored checklist and channel-specific drafts in ten minutesa spreadsheet of every playlist in one click
Aha momentFirst moment of valuepasting a repo URL and seeing the three fixes that matterthe exported file opening with every song listed
Why nowWhat changed recently?AI agents can now do the drafting, but don't know community rulesmore people switching between music apps
ProofWhy believe it?public rubric; rules link to each platform's own guidelinesworks on the five biggest platforms (list them)
ContrastWhy not the alternative? (dated)generic marketing prompt packs don't know V2EX or Show HN etiquette (2026-10)<tool X> doesn't support Chinese platforms (checked 2026-10)
CategoryWhat shelf does it sit on?launch checklist / developer marketingplaylist backup tool

Tests for a one-liner

  • Stranger test: would someone who's never heard of it understand it out of context, e.g. in a list of post titles?
  • Swap test: if you swap in a competitor's name, is it still true? If yes, it's not a position. In a crowded category where any plain outcome line fails this, let the bullets carry the difference — and say so.
  • Breath test: can it be said aloud in one breath? (≈ 12 English words or ≈ 20 Chinese characters; Latin brand names inside Chinese count as one character each.)
  • No-buzzword test: remove "AI-powered", "blazing", "next-gen", "revolutionary", 「神器」「颠覆」. Does it still say something?
  • Founder-voice test: every sentence in the author's voice ("I built it because…") is the author's own words or a [TODO].
  • Claims test: nothing in it appears under Claims we won't make.

Template for .megaphone/brief.md

Write it in the user's language. Include the Chinese one-liner only if a Chinese channel or README is planned (and the English one only if an English channel is).

# Launch brief — <project>

**One-liner (EN):** …
**一句话(中文):** …
**Alternatives considered:** … / … (and why the chosen one wins)

| | |
|---|---|
| Audience | … |
| Trigger | … |
| Today they use | … |
| Promise | … |
| Aha moment | … |
| Why now | … |
| Proof | … |
| Contrast (as of <date>) | … |
| Category | … |

## Origin story (author's words only)
[TODO: ask the author] / "<quote>"

## Three bullets (for README and posts)
- **<benefit>** — <how>
- …

## Words to use / avoid
- Use: <words the audience itself uses — from issues, forum threads, reviews>
- Avoid: <jargon, buzzwords, internal names>

## Claims we won't make
- Legal or safety guarantees ("legal", "zero risk", "official", "零风险")
- Wording that reads as circumvention ("bypass", "破解", "突破限制", "抓包")
- Uses that invite misuse (e.g. looking up other people's data)
- <anything from audit.md → Risk & readiness>

## Standard disclaimer (for README and posts, if the project touches other platforms)
"<Project> is not affiliated with <platforms>. It <scope>; it does not <what it doesn't do>."

## Open questions
- [TODO] …

Write the Chinese one-liner natively, not as a translation: Chinese audiences respond to concrete pain ("再也不用手动…") and a clear 「适合谁」.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.