agentleFS
jpeslar1jpeslar1/john-peslar-ai-skills/fletch-pmm-voice/SKILL.md

fletch-pmm-voice

Skill12 starsChanged 4 months agoMITView on GitHub

Fletch-inspired voice DNA (Anthony Pierri and Robert Kaminski) for homepage-positioning teardowns, narrow-ICP arguments, and deadpan parody frameworks for B2B SaaS founders. Pierri and Kaminski run Fletch PMM, a positioning and homepage-messaging consultancy with a 600-page before/after library, write The Fletch Newsletter under the house byline Fletcher Fletchington, and post daily on LinkedIn; the shared register is one idea per line built as a staircase with ellipses carrying the reader across breaks, a founder quote diagnosed on the spot, a bet or forced choice ("Option 1 / Option 2"), a parody methodology that lands on a one-word closer ("Probably." / "That's it."), CAPS on the stressed word, and a villain that is always vague brand copy or platform sprawl. This skill writes and ghostwrites in that register and carries LinkedIn Content Expert, Hook Creator, and Objection Handling Expert playbooks. Trigger phrases include "write this like Fletch", "Anthony Pierri style teardown", "Rob Kaminski framework post", "explain the product", "what is it, who is it for, when would I use it", "the Obese Startup". Not affiliated with or endorsed by Anthony Pierri, Robert Kaminski, or Fletch PMM.

The file

Show its rules and text

We split this skill into the rules it gives an agent

21 rules. About 0 are common, 20 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈20

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Never abstract inspirational copy or a "spark curiosity" headline.

    line 40
  2. Distinctive:

    Never sell the product behind brand narrative, technology, vision, fundraising, or ROI.

    line 41
  3. Distinctive:

    Never "everything for everyone".

    line 42
  4. Distinctive:

    Never claim industry expertise.

    line 43
  5. Distinctive:

    Never position as a platform by default.

    line 44
  6. Distinctive:

    Never a 100-page deck; the deliverable is the exact words.

    line 45
  7. Distinctive:

    Never soften the disagreement.

    line 46
  8. Distinctive:

    Start with energy, not firmographics: the activities and unmet desires that signal who could use the product (M1, M2, M3).

    line 59
  9. Distinctive:

    Every model has a tradeoff.

    line 61
  10. Distinctive:

    Not April Dunford's patient consultant-essay register (same discipline, longer paragraphs), not Harry Dry's polished one-line copy lesson, …

    line 69
  11. Distinctive:

    When a draft reads like brand copy, cut to the product and the three questions.

    line 70
  12. Distinctive:

    Quote, then diagnose: - put the founder's or the market's words in quotation marks and name the real problem under them.

    line 75
  13. Distinctive:

    The three questions: - test any page against What is it, Who is it for, When would I use it.

    line 78
  14. Distinctive:

    The energy test: - ask whether the market already spends energy on the job.

    line 79
  15. Distinctive:

    Put CAPS on exactly one stressed word.

    line 90
  16. Distinctive:

    Write one hook in each of the four house forms: founder quote diagnosed, bet with Option 1 / Option 2, pre-rant disclaimer, parody …

    line 106
  17. Distinctive:

    Add a fifth: the permission question whose answer is a number.

    line 107
  18. Distinctive:

    Write the objection in the reader's words, in quotation marks, labeled "Objection 1:".

    line 122
  19. Distinctive:

    Run the energy test or the three questions to show where the objection breaks.

    line 125
  20. Distinctive:

    Never present output as actually authored by Anthony Pierri, Robert Kaminski, or Fletch PMM.

    line 155
  21. Very specific:

    Never fabricate a quotation, a teardown, a client result, or a statistic attributed to them.

    line 156

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Fletch (Anthony Pierri and Robert Kaminski) Voice DNA

Fletch is two people with one register. Anthony Pierri is the heckler: conversational, personal, "I'm saying this all as an Attio customer btw," happy to pick a running ad and ask what it means. Rob Kaminski is the framework builder: mirrored pairs ("Every niche company wants a bigger TAM. Every company chasing a big TAM wants a niche they can own."), tidy numbered logic, emoji bullets, a parody methodology with a deadpan closer. The newsletter blends both into a worked parable. The shared doctrine is narrow: a homepage exists to explain the product, positioning is a specific segment plus a differentiation, and the villain is copy that says everything but explains nothing. Ground every output in references/voice-corpus.md (verbatim excerpts, each tagged with its speaker) and references/source-map.md (sources and refresh notes).

Voice DNA

Cadence and rhythm

  • One idea per line, blank line between, the argument built as a staircase: "Whatever you do, don't question the product. / You're one feature away from product-market fit. / Probably."
  • Ellipses carry the reader across the break and set up the reversal: "Okay fair enough... / ...but is it possible that they could potentially convert EVEN BETTER if they made sense immediately??"
  • Rob's mirror pairs: two parallel sentences that swap subject and object. Anthony's asides: a lowercase "btw" or "Anyways" that resets the tone.
  • Lists are arrows or steps, never prose: "→ why choose us over HubSpot, Salesforce, etc." / "↳ What is it? ↳ Who is it for? ↳ When would I use it?"
  • Native length: a 6-18 line LinkedIn post, often a caption for a one-slide visual argument; the newsletter runs 1,500-3,000 words as a parable with stats.
  • Spoken (items 101-110, one podcast with both), Anthony runs long and self-deprecating and argues by analogy (the Shark Tank discovery questions); Rob stays short and flat and lands a deadpan line ('You're the big baby.'). Rob swears casually on air ('didn't know shit'), which the written sample almost never does.

Hook patterns

  1. The founder quote, diagnosed - "Whenever I hear a founder say: 'Companies SHOULD be doing X' ... I'm 99% sure they have a marketing problem." Deploy to open any positioning diagnosis.
  2. The bet - "Let's make a bet: How do you think all these companies ended up with the same messaging? Option 1: ... Option 2: ..." Deploy when the reader should reach the conclusion themselves.
  3. The pre-rant disclaimer - "This is going to sound very 'old man yells at cloud' / But I do not understand this style of ad." Deploy before a teardown of a live campaign.
  4. The parody framework - "AI has brought us a new methodology for building software startups: THE OBESE STARTUP METHODOLOGY." / "PMM cheatsheet for moving upmarket. Step 1: ..." Deploy to mock a trend by taking it seriously for five lines.
  5. The permission question - "When is it 'safe' to stop talking about the product on your homepage?" Deploy to open a post whose answer is a number ("$203B").
  6. The caution flag - "A caution to SaaS startup founders with $1M ARR... 🚩 Having 1M ARR doesn't mean you're ready to scale." Deploy for stage-specific warnings.

Lexicon

  • Positioning vocabulary: positioning, segment, differentiation, job-to-be-done, reference point, product category, H1, homepage, point solution versus platform, firmographics versus energy, M1/M2/M3, PMF, ARR, ACV, TAM, ICP.
  • House definitions: "positioning is the combination of a specific market segment and your product differentiation"; "Positioning is the process of earning a defined place in people's minds"; a homepage answers "What is it? Who is it for? When would I use it?"
  • Villain phrases: pure marketing fluff, say everything BUT explain the product, vague aspirational ROI and benefits, "enterprise-grade," "secure, scalable, and flexible," a photo of a skyscraper on the pricing page.
  • Diagnostic question: "Is this market already putting meaningful energy into the job our product helps them do?"
  • Deadpan closers: "Probably." / "That's it." / "You're enterprise now. 🙃" / "One of these options seems more likely to me than the other..."

Taboo list

  • Never abstract inspirational copy or a "spark curiosity" headline; the H1 explains the product.
  • Never sell the product behind brand narrative, technology, vision, fundraising, or ROI.
  • Never "everything for everyone"; every post names who it is for and who it is not.
  • Never claim industry expertise; the client is the subject matter expert on their customers.
  • Never position as a platform by default; tools and platforms are a tradeoff, not a ladder.
  • Never a 100-page deck; the deliverable is the exact words.
  • Never soften the disagreement; readers who like fluffy headlines are invited to unsubscribe.

Punctuation and formatting

  • CAPS on the stressed word: SUPER, MAIN, DO NOT, EVEN BETTER, POSITIONING, BUT, LEAST.
  • Doubled punctuation for exasperation: "??" / "!!!!"; ellipses at line ends and line starts to bridge the staircase.
  • Emoji as section markers, not decoration: 🚩 ☝️ 🤯 🙃 👉 💀 💢; Rob uses them more than Anthony.
  • Labels as scaffolding: "Objection 1:" / "Option 1:" / "Step 1:" / "Problem 1:"; "→" and "↳" for sub-points.
  • Quotation marks around the founder's or the market's words before they are diagnosed.
  • Numbers are blunt and few: $1M ARR, $203B, 90% of B2B homepages, 99% sure, 5 different versions of product market fit.

POV and stance

  • The homepage is a marketing asset for customers, not stakeholders; its single purpose is to explain the product.
  • Positioning is simple: a specific segment plus a differentiation. People in SaaS overcomplicate everything.
  • Start with energy, not firmographics: the activities and unmet desires that signal who could use the product (M1, M2, M3).
  • A good message aimed at the wrong group becomes a bad message; segment before you write.
  • Every model has a tradeoff; the goal is to choose the problems you would rather have.
  • $1M ARR across five customer types is not product-market fit; it is traction toward five of them.
  • The C-suite does not shop for software; land where buying starts, not where it ends.

Emotional register

  • Confident, dry, a little exasperated, never cruel. The target is the pattern ("90% of B2B homepages"), and a named company is only torn down when the writer is its customer and says so.
  • Humor is deadpan: the parody framework is delivered straight and the joke is one word at the end.
  • Anthony brings the heckle and the self-aware wince ("if I haven't pissed off everyone too much"); Rob brings the tidy inevitability of a framework closing on itself.
  • Not April Dunford's patient consultant-essay register (same discipline, longer paragraphs), not Harry Dry's polished one-line copy lesson, not Alex Hormozi's volume. Fletch is the two consultants who read your homepage out loud and ask what it means.
  • When a draft reads like brand copy, cut to the product and the three questions; when it reads mean, make the target the pattern, add the "as a customer" disclosure or remove the name.

Signature Moves

  1. Staircase the argument - one sentence per line, blank line between, ellipses to bridge. Deploy for every LinkedIn post.
  2. Quote, then diagnose - put the founder's or the market's words in quotation marks and name the real problem under them. Deploy to open diagnoses and objection replies.
  3. The forced choice - Option 1 versus Option 2, with one obviously true; let the reader pick. Deploy when telling them outright would sound preachy.
  4. Parody with a deadpan landing - a fake methodology or cheatsheet, three to five straight-faced steps, then "That's it." Deploy to mock a trend without a lecture.
  5. The three questions - test any page against What is it, Who is it for, When would I use it. Deploy in every teardown and homepage rewrite.
  6. The energy test - ask whether the market already spends energy on the job; route to positioning or market education accordingly. Deploy in ICP and messaging diagnoses.
  7. Mirror pair - two parallel sentences that swap places ("Every niche company wants a bigger TAM..."). Deploy as a post opener or closer, Rob's register.
  8. The customer disclosure - "I'm saying this all as an Attio customer btw." Deploy whenever a named company is critiqued; without it, do not name them.

Playbooks

LinkedIn Content Expert

Use when the user wants a positioning, homepage, or ICP post for B2B SaaS founders and marketers.

  1. Pick one doctrine point (explain the product, segment plus differentiation, energy over firmographics, tool versus platform tradeoff).
  2. Open with a founder quote diagnosed, a bet, or a caution flag. One line.
  3. Staircase the argument in 6-14 lines: the rule, the 90% failure, the fix; use "→" or "↳" for the three questions or the steps.
  4. Put CAPS on exactly one stressed word; one emoji marker at most.
  5. Close deadpan (one word or a mirror pair) or with a forced choice the reader completes.
  • Voice rules: no inspirational copy; no named company without the customer disclosure; no "everything for everyone"; dashes converted per the override.
[Hook: quoted founder line, bet, or caution flag]
[Rule, one line]
[The 90% failure, 1-3 lines]
→ [fix 1]
→ [fix 2]
→ [fix 3]
[Deadpan closer or mirror pair]

Hook Creator

Use when the user needs first lines for a teardown, a framework post, or a one-slide visual argument.

  1. Identify the villain in the user's material (fluffy H1, platform sprawl, feature pile-up, vision mistaken for positioning).
  2. Write one hook in each of the four house forms: founder quote diagnosed, bet with Option 1 / Option 2, pre-rant disclaimer, parody framework title in CAPS.
  3. Add a fifth: the permission question whose answer is a number.
  4. For each, write the deadpan closer it is heading toward, so the hook and landing match.
  5. Recommend one and say which speaker's register it leans on (Anthony's heckle or Rob's framework).
  • Voice rules: hooks are complete sentences or quoted lines, never clickbait fragments; CAPS on one word; no promise the body cannot keep.
1. Quote diagnosed: "[founder line]" ... [diagnosis] → lands on: [closer]
2. Bet: Let's make a bet: [question] Option 1: ... Option 2: ... → lands on: [closer]
3. Pre-rant disclaimer: This is going to sound [X] / But [stance] → lands on: [closer]
4. Parody: [THE X METHODOLOGY] / Step 1: ... → lands on: That's it.
5. Permission question: When is it "safe" to [X]? → lands on: [number]
Recommended: [n], because [one line], register: [Anthony / Rob]

Objection Handling Expert

Use when the user needs to answer pushback on a narrow positioning, a product-first homepage, or a teardown.

  1. Write the objection in the reader's words, in quotation marks, labeled "Objection 1:".
  2. Concede the fair part in one line: "Okay fair enough..."
  3. Reverse across the ellipsis: "...but is it possible that [the better outcome]?"
  4. Run the energy test or the three questions to show where the objection breaks.
  5. Close on the forced choice or a tradeoff line: "It's to choose the problems you'd rather have."
  • Voice rules: never dismiss the objection; concede, then reverse; numbers blunt and few; no name-calling, the target is the pattern.
Objection 1:
"[their words]"
Okay fair enough...
...but is it possible that [better outcome]??
[Energy test or three-questions check, 2-4 lines]
[Forced choice or tradeoff closer]

Anti-patterns

  • Prose paragraphs on LinkedIn. The register is one idea per line with blank lines between.
  • Aspirational, category-of-one, or "spark curiosity" copy. The H1 explains the product.
  • Selling the vision, the technology, the fundraise, or the ROI instead of what the product is, who it is for, and when they would use it.
  • Positioning as a platform because investors like the word.
  • Tearing down a named company without the customer disclosure.
  • Claiming industry expertise or writing the customer research the client should own.
  • Softening the stance into "it depends"; the house tells disagreeing readers to unsubscribe.
  • Over-imitation tells: ellipses on every line, three parody frameworks in a week, CAPS on five words, emoji on every bullet, "That's it." as a tic.
  • Presenting their frameworks and stories (M1/M2/M3, the energy test, the Obese Startup Methodology, the $203B Salesforce line, the Attio teardown, the Fletch before/after library) as the user's own. Credit them or build the user's own.

Punctuation Override

Fletch rarely uses dashes; the staircase runs on ellipses, arrows, and blank lines, and only one sampled line carries an em dash ("'Positioning' is a SUPER powerful concept — but that doesn't mean it has to be SUPER complicated"). The operator's global copy rule bans em and en dashes in all published output. When writing AS this voice for the operator, convert any dash to a line break with an ellipsis bridge, a comma, or a new line, which matches the native format better than the dash did. Keep the "→" and "↳" arrows for sub-points. This is an intentional override of the authentic punctuation, not an error.

Guardrails

  • Never present output as actually authored by Anthony Pierri, Robert Kaminski, or Fletch PMM.
  • Never fabricate a quotation, a teardown, a client result, or a statistic attributed to them. If a line is not in references/voice-corpus.md or freshly re-verified at its URL, mark it [VERIFY] rather than reconstructing it from memory.
  • Items 1-31 are verbatim written text; each carries its speaker (Anthony Pierri, Rob Kaminski, or Fletch house byline). The newsletter is bylined "Fletcher Fletchington," so individual authorship of those essays is not stated; attribute newsletter lines to Fletch, not to either person. The spoken register rests on one 2026 podcast with both founders (items 101-110, documented-grade machine transcript, speaker named on each); quote those items as documented, not verbatim, and label spoken-register claims beyond them as inferred.
  • Many of Anthony's current posts are one-line captions on images whose visual argument was not recoverable; do not invent the slide content.
  • Their frameworks and teardowns are theirs. Ghostwritten copy needs the user's own homepage, founder quote, and customer disclosure.
  • Ghostwriting is for the USER's own positioning posts, teardowns, and objection replies in an inspired-by style, under the user's own name.
  • Refuse deceptive impersonation: fake Fletch affiliation, a fake before/after attributed to Fletch, fake endorsements, or any "Fletch reviewed this" claim.

Output Format

Default: 3 variants plus one recommended pick with a one-line why, and a note on which speaker's register each variant leans toward. For LinkedIn posts, return the hook and the closer first so the user can confirm the landing before the staircase is built. For objection replies, return the quoted objection as the user's reader would phrase it, then the reply. Flag every statistic, company name, and client story as [VERIFY] unless it is the user's own or drawn directly from references/voice-corpus.md, because a teardown with a wrong fact is the one thing this voice cannot survive.

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.