Skip to content

cgallic/kai-cmo-harness / harness/voice-gate

voice-gate

Verified

Line-editor pass on any content draft against a client voice guide. Flags cadence issues, signature-move overuse, voice-guide violations, receipt gaps — things a rule-based gate can't catch. Use after /content-gate and before human review.

FileKindSize
README.mdmarkdown3.8 KB
SKILL.mdskill md8.8 KB
TUNING-LOG.mdmarkdown3.3 KB
examples/example-blog-voice-gate-report.mdmarkdown27.5 KB
judge-prompt.mdmarkdown13.5 KB
run-log.mdmarkdown242 B

examples/example-blog-voice-gate-report.md

View on GitHub
# Voice Gate Report — how-to-write-a-podcast-pitch.md

**Reviewer:** /voice-gate (LLM-as-judge, simulated run)
**Calibration:** lexie-smith-writing-guide.md
**Target file:** clients/GrowthModeOn/outputs/blog-posts/how-to-write-a-podcast-pitch.md
**Word count:** ~1,670 body words
**Total issues:** 20
**Severity breakdown:** High 6, Medium 9, Low 5
**Overall verdict:** FAIL — The tactical spine (Pre Pitch Op acronym + customization flow + four-part pitch anatomy) is sound and earns its place. The issues are at the sentence-cadence layer: the "I repeat" signature move fires 3× where the voice guide caps it at 1×, triple-list cadence has become the dominant rhythm, and two of the stronger "receipts" in the post are hedged rather than named. Single biggest fix: rewrite every third triple-list into a different sentence shape in one pass.

---

## Summary Table

| # | Severity | Category | Location | Issue |
|---|---|---|---|---|
| 1 | High | Voice-guide violation | Lines 58, 100, 112 | "I repeat" / bolded-repeat fires 3× — voice guide §1.2 caps at 1× |
| 2 | High | Sing-songy rhythm / cadence monotony | Line 21 (intro) | Three of four intro sentences end on triple-lists; meta beat delays the payload |
| 3 | High | Sing-songy rhythm / cadence monotony | Lines 31–34 | All four bullets open "You are pitching X, not Y" — same shape four times |
| 4 | High | Repetition / redundancy | Line 100 | "Do not attach files" idea restated 5× in 6 lines |
| 5 | High | Receipt gaps | Line 100 | "Per my own HARO receipts" — self-quoted, no year, no source — hedged receipt |
| 6 | High | Meta-commentary breaks flow | Lines 21 + 128 | "That is the short version" + "Let me recap" — two narrator interjections, meta-beat arrives before reader earns a pause |
| 7 | Medium | Algorithmic Authorship violations | Lines 34, 52–56 | Condition before main clause on line 34; five research bullets are noun-phrases where verb-first imperatives extract better |
| 8 | Medium | Voice-guide violation | Lines 66–74 | Pre Pitch Op acronym uses em-dashes as letter-to-definition separators 10× — §11 DON'T #11 "em-dash crutch" |
| 9 | Medium | Sing-songy rhythm / cadence monotony | Lines 84–87 | Three-pitch example + closer stacks five triple-beats in a row |
| 10 | Medium | Sing-songy rhythm / cadence monotony | FAQ section (lines 136–155) | Six of seven FAQ answers close on the same contrastive "X, not Y" shape |
| 11 | Medium | Receipt gaps | Line 105 | "I had a client recently land a segment on New York Live" — no year, no named client, no number |
| 12 | Medium | FAQ atomicity | Line 146 | "How many podcasts…" answer makes 3 claims + drops an internal link inside the FAQ |
| 13 | Medium | Repetition / redundancy | Lines 78–87 | "Same client, same spring-break campaign" stated on line 81 then re-said on line 87 |
| 14 | Medium | Repetition / redundancy | Section 1 (lines 25–36) | "pitch" / "pitching" fires 14× in 12 lines |
| 15 | Medium | Repetition / redundancy | Section 2 (lines 47–58) | "research" / "researched" / "research turns up" / "do the research" appears 5× in 12 lines |
| 16 | Low | Other editorial catches | Line 21 | "could not get into" is flatter than §12.1 voice-guide anchor "couldn't get a foot in the door of" |
| 17 | Low | Meta-commentary breaks flow | Line 41 | "Work through these in order. Do not skip." — teacher-ish admonishment |
| 18 | Low | Other editorial catches | Line 118 | "40 other pitches in the inbox today" — invented placeholder number, voice guide §11 DO #4 demands a receipt |
| 19 | Low | Other editorial catches | Line 39 | "This is the exact workflow my team uses" — "uses" is a weak verb; §12.5 example has more bite |
| 20 | Low | Transitions and flow | Line 89 | "Write the pitch itself. Four parts, no filler." drops in as a list header with no transitional beat |

---

## Detailed Issues

### Issue 1 — "I repeat" / bolded-repeat fires 3× in one post
**Severity:** High
**Category:** Voice-guide violation
**Location:** Lines 58, 100, 112
**Voice-guide rule:** §1.2 "Use the literal phrase 'I repeat' max once per post." Also §2.2 translation rule: "**Bold** the second occurrence. Let the typography do the work."
**Current text (verbatim):**
> Line 58: "**Do not force a fit.** That is one of my repeated rules in PR. **Do not force a fit.**"
> Line 100: "Do not attach files. I repeat. **Do not attach files.**"
> Line 112: "Spell it correctly. I repeat. **Spell it correctly.**"

**Why it's an issue:** The voice guide caps the "I repeat" signature move at 1× per post. The post fires it three times — once with the literal phrase on line 100, once as a paraphrased "one of my repeated rules" bolded-repeat on line 58, and a third literal "I repeat" on line 112. When a signature move fires every ~40 lines the emphasis stops emphasizing and the pattern reads as a crutch rather than a signature.

**Recommended fix:** Pick the single strongest occurrence — the line 112 "Spell it correctly" beat has the pet-peeve "Hi Lexi" payoff and earns its place. Replace the other two with different emphasis patterns from §8 — a short sentence on its own line, a "Period." close, or a blockquote.

---

### Issue 2 — Sing-songy intro paragraph
**Severity:** High
**Category:** Sing-songy rhythm / cadence monotony
**Location:** Line 21 (opening paragraph)
**Voice-guide rule:** §3 Sentence Architecture ("Mostly short to medium… punchy short closers") + §2.4 cold-open rule + Editorial: triple-list cadence monotony (not explicitly in the voice guide but implicit in §3 and §8).
**Current text (verbatim):**
> A podcast pitch that books the show is short, personalized, timely, and does the host's research for them. It names one specific episode you listened to, answers the question "why you, why now, why this audience" in two or three sentences, and proves you can deliver with a credential, a clip, or a recent result. That is the short version. Now let me show you how that actually plays out, because in 15+ years of PR I have sent thousands of pitches, killed plenty of bad ones, and watched a few sentences land my clients on shows their competitors could not get into.

**Why it's an issue:** Sentence 1 lists four attributes (short / personalized / timely / does-the-research). Sentence 2 re-lists three of them as verbs (names / answers / proves). Sentence 4 ends on another triple-list (sent / killed / watched). Three of the four sentences end on three-item lists. The cadence becomes predictable by sentence two. "That is the short version" is a meta-beat that delays the payload just as the reader is leaning in.

**Recommended fix:** Cut "That is the short version. Now let me show you how that actually plays out." Merge the "15+ years" receipt into sentence 2 as evidence for the definition. Break at least one of the remaining triple-lists into a different sentence shape — a short knockout, a conditional, or a named-client receipt.

---

### Issue 3 — Four-bullet "You are pitching X, not Y" monotony
**Severity:** High
**Category:** Sing-songy rhythm / cadence monotony
**Location:** Lines 31–34 ("Here is what that means for your pitch")
**Voice-guide rule:** Editorial: identical-shape cadence monotony. §3 Sentence Architecture and §8 Emphasis Patterns both warn against repeated sentence shapes.
**Current text (verbatim):**
> - You are pitching a conversation, not a press release. Skip the announcement energy.
> - You are pitching the host's audience, not the host. Prove you know who listens.
> - You are pitching a format, not a generic slot. Solo expert? Co-host debate? Case study interview? Match the format.
> - You are pitching timing, not just a topic. A host who just released an episode on your exact subject does not need another one next week.

**Why it's an issue:** All four bullets open "You are pitching X, not Y." The rhetorical move is strong once; four in a row becomes a drumbeat the reader tunes out by bullet two. The concept in each bullet is distinct — the delivery flattens them into one mold.

**Recommended fix:** Keep bullet 1 as the anchor pattern. Rewrite bullets 2–4 into different structures — a question, a short imperative, a named-client receipt, a conditional.

---

### Issue 4 — "Attach nothing" repetition stack
**Severity:** High
**Category:** Repetition / redundancy
**Location:** Line 100 (Section 6, "Attach nothing. Link to everything.")
**Voice-guide rule:** §1.2 / §2.2 — the bolded-repeat pattern is capped; §2.2 says "Let the typography do the work."
**Current text (verbatim):**
> Do not attach files. I repeat. **Do not attach files.** Per my own HARO receipts, "the attachment you sent gets dropped between your inbox and their inbox." Hosts filter by spam risk. Your PDF is spam-risk. A clean link to a one-page Notion doc or a Google Doc is better in every way.

**Why it's an issue:** The same idea ("don't attach files") is stated five separate times in six lines: (1) "Do not attach files," (2) "I repeat," (3) bolded "Do not attach files," (4) the HARO self-quote, (5) "Your PDF is spam-risk." By the third repetition the reader is ahead of the sentence.

**Recommended fix:** Keep the bolded "Do not attach files." + one clean reason + the alternative (link to Notion / Google Doc). Cut the HARO self-quote or swap it for a specific year + platform anecdote from the persona file. Collapse to three sentences max.

---

### Issue 5 — "Per my own HARO receipts" is not a receipt
**Severity:** High
**Category:** Receipt gaps
**Location:** Line 100
**Voice-guide rule:** §6 Receipts Pattern + §11 DO #4: "Back every claim with a receipt — number, year, client outcome, anecdote, or industry stat." §6 lists seven receipt types — "self-quoted paraphrase" is not one of them.
**Current text (verbatim):**
> Per my own HARO receipts, "the attachment you sent gets dropped between your inbox and their inbox."

**Why it's an issue:** "My own HARO receipts" is a hedge, not a receipt. No year. No platform. No number. The quote marks around a self-paraphrase read as invented. The persona file §Receipts lists multiple named-client HARO anecdotes that would carry the weight this line is reaching for.

**Recommended fix:** Swap for a named year + platform receipt (e.g., "When I audited my 2023 HARO queue…") or pull a specific named-client anecdote from the persona file. Cut the self-quote marks either way.

---

### Issue 6 — Meta-commentary in the intro and recap header
**Severity:** High
**Category:** Meta-commentary breaks flow
**Location:** Line 21 ("That is the short version. Now let me show you how that actually plays out") + Line 128 ("Let me recap" as H2)
**Voice-guide rule:** §2.4 Blog Requirements + §5.5 SEO Blog Pattern. Meta-commentary pairs are allowed but the payload must come first. §11 DON'T #2 forbids "In conclusion / In summary" and §3 emphasizes punchy closers over narrator interjections.
**Current text (verbatim):**
> Line 21: "That is the short version. Now let me show you how that actually plays out"
> Line 128: "Let me recap" (used as an H2 section heading)

**Why it's an issue:** "Let me recap" as an H2 is voice-accurate on its own. Paired with the intro meta-beat, the post has two narrator-interrupts-the-teaching moments — plus "Work through these in order. Do not skip" on line 41 makes a third. The intro meta specifically delays the 15+ years receipt that should be the hook.

**Recommended fix:** Lead with the receipt, not the meta. Cut "That is the short version. Now let me show you how that actually plays out" from the intro. Keep "Let me recap" as the H2 — just not both.

---

### Issue 7 — Algorithmic Authorship violations
**Severity:** Medium
**Category:** Algorithmic Authorship violations
**Location:** Lines 34, 52–56
**Voice-guide rule:** §2.4 "Run the Algorithmic Authorship pass AFTER the voice pass" — two AA rules are not clean.
**Current text (verbatim):**
> Line 34: "A host who just released an episode on your exact subject does not need another one next week."
> Lines 52–56: "Recent guests. Who has the host already featured this quarter? / Episode topics in the last 90 days. Do not pitch a subject they just covered. / Their solo vs. interview ratio. Some shows are 100% interviews. Others are 80% solo. / Format clues in the episode titles…"

**Why it's an issue:** Two AA rules violated: (1) condition before main clause on line 34 — AA rule is "Do X if Y," not "If Y, do X"; (2) five research-target bullets are noun-phrases rather than verb-first imperatives. AI Overview extraction prefers verb-led imperatives for instructional lists.

**Recommended fix:** Rewrite line 34 as "Skip the show if the host just released an episode on your exact subject." Convert the five research bullets to verb-first: "Check recent guests. / Scan episode topics from the last 90 days. / Map the solo-to-interview ratio…"

---

### Issue 8 — Pre Pitch Op acronym uses em-dashes 10×
**Severity:** Medium
**Category:** Voice-guide violation
**Location:** Lines 66–74 (the Pre Pitch Op bulleted acronym)
**Voice-guide rule:** §11 DON'T #11: "Don't use em-dashes as a stylistic crutch. She rarely uses them."
**Current text (verbatim):**
> - **P — Proximity** (location: does this matter to the host's audience geographically?)
> - **R — Relevance** (does this fit the show's actual niche?)
> - **E — Extremeness** (is there something remarkable about the scale, speed, or stakes?)
> (pattern repeats for P, I, T, C, H, O, P — 10 em-dashes total in one block)

**Why it's an issue:** The voice guide flags em-dashes as a stylistic crutch. The acronym block uses em-dashes as letter-to-definition separators 10 times. This is a known carve-out from prior drafts (the session log acknowledges "em-dashes reduced to 10 per file (all functional)"), but the guide does not exempt acronym separators — so this is a voice-guide violation waiting on a documented exception.

**Recommended fix:** Two clean options — swap to colons ("**P: Proximity**") or a period-space ("**P.** Proximity"). Either reads cleaner to the voice and avoids the guide violation. If the carve-out stands, document it in the voice guide.

---

### Issue 9 — Three-pitch example stacks five triple-beats
**Severity:** Medium
**Category:** Sing-songy rhythm / cadence monotony
**Location:** Lines 84–87 (customization step example)
**Voice-guide rule:** Editorial: triple-list cadence monotony. §3 Sentence Architecture.
**Current text (verbatim):**
> - Parenting podcast: "Three things moms overpack for spring break (and what to bring instead)."
> - Travel-tech podcast: "How mobile booking behavior shifts 28% in the two weeks before spring break."
> - Women-in-business podcast: "The founder story of a travel company built by a mom for moms."
>
> Same topic. Same client. Three completely different pitches. Each one matches the host's format, audience, and last three episodes.

**Why it's an issue:** Three bullets of the same shape, then three short sentences of the same shape ("Same topic. Same client. Three completely different pitches."), then a sentence that ends on another triple-list ("format, audience, and last three episodes"). Five triple-beats stacked in six lines. The example itself is excellent — the closing frame is what drags.

**Recommended fix:** Keep the three-bullet example (genuinely the teaching moment of the post). Cut "Same topic. Same client. Three completely different pitches." or collapse to one sentence.

---

### Issue 10 — FAQ close-cadence uniformity
**Severity:** Medium
**Category:** Sing-songy rhythm / cadence monotony
**Location:** FAQ section (lines 136–155)
**Voice-guide rule:** §3 Sentence Architecture + §8 Emphasis Patterns — signature moves lose power when they become the only move.
**Current text (verbatim):**
> "Long pitches get archived, not read."
> "Media kits matter for brand deals, not for podcast guesting."
> "That reads as lazy, not persistent."
> "The research-based sentence has to come from you."
> "That is the whole job."

**Why it's an issue:** The "X, not Y" contrastive close (or its cousin, the one-sentence knockout) is a signature move — and six of seven FAQs end on the same shape. When every FAQ ends the same way the move stops working.

**Recommended fix:** Keep 2–3 FAQs closing on the contrastive beat. Rewrite 3–4 to close with a different shape — a question, a conditional, a named receipt, or a short imperative.

---

### Issue 11 — "New York Live" segment receipt is vague
**Severity:** Medium
**Category:** Receipt gaps
**Location:** Line 105
**Voice-guide rule:** §6 Receipts Pattern + §11 DO #4.
**Current text (verbatim):**
> "I had a client recently land a segment on New York Live because she did a single follow-up with a new timeliness hook."

**Why it's an issue:** Compare to the voice-guide-approved receipt in §12.3: "Sunday covers up more than 70%. Sunday profits up more than 150%." Named venue, specific numbers, named partner. This one has "recently" for the year, no client name, no number, no named host.

**Recommended fix:** Either cite year + segment title + host name + hook used, or swap for a named-client anecdote from the persona file's receipt library that has the specifics.

---

### Issue 12 — FAQ atomicity breaks on "How many podcasts…"
**Severity:** Medium
**Category:** FAQ atomicity
**Location:** Line 146 ("How many podcasts should I pitch at a time?")
**Voice-guide rule:** §2.4 Blog Requirements — FAQ answers 40–60 words, atomic Q+A.
**Current text (verbatim):**
> Quality over volume. Ten well-researched, personalized pitches beat 100 generic ones every time. The whole point of [a founder's personal brand](/personal-branding/) is to be the only option, not one of a hundred. A batch of 10, sent Tuesday to Thursday, with individual follow-ups scheduled is a realistic weekly cadence.

**Why it's an issue:** Three separate claims in one answer: (a) quality over volume, (b) pillar-link tangent about personal brand as "the only option," (c) batch cadence of 10 Tue–Thu. For FAQPage schema and Featured Snippet extraction, one claim per question is cleaner. The internal link inside the FAQ answer is also unusual — internal links typically live in body copy, not FAQ answers, because they can interfere with schema-extracted snippet display.

**Recommended fix:** Cut the "whole point of personal brand" sentence + link. Answer reduces to two clean claims: volume target + cadence.

---

### Issue 13 — "Same client, same campaign" frame stated twice
**Severity:** Medium
**Category:** Repetition / redundancy
**Location:** Lines 78–87
**Voice-guide rule:** §3 Sentence Architecture — "Strip run-ons and redundancy in writing."
**Current text (verbatim):**
> Line 81: "Example from my agency. Same client, same spring-break campaign, three different podcast pitches:"
> Line 87: "Same topic. Same client. Three completely different pitches."

**Why it's an issue:** The "same client / same campaign" frame is stated on line 81, then re-said on line 87 as a summary beat. One of the two is redundant — the second one is also where the five-triple-beat stack in Issue 9 starts.

**Recommended fix:** Keep line 81 as the setup. Cut or reshape line 87's closer. Pairs well with the Issue 9 fix.

---

### Issue 14 — "Pitch" / "pitching" fires 14× in 12 lines
**Severity:** Medium
**Category:** Repetition / redundancy
**Location:** Section 1 (lines 25–36)
**Voice-guide rule:** §3 Sentence Architecture — break patterns, no word-stacking.
**Current text (verbatim):**
> Lines 25–36 contain "pitch / pitching / pitches" 14 times across 12 lines, e.g.:
> "A media pitch sells a story to a journalist…"
> "A podcast pitch sells a conversation…"
> "Here is what that means for your pitch"
> "You are pitching a conversation, not a press release."
> (pattern continues through the four-bullet block in Issue 3)

**Why it's an issue:** Some repetition is necessary — "pitch" is the post topic. By line 34 the reader is pattern-fatigued. The density is highest in the four-bullet block, which Issue 3 also flags.

**Recommended fix:** The four-bullet rewrite in Issue 3 will absorb most of this. For the rest, swap in "send," "submit," "propose," "open with" where the meaning allows. Don't strip every instance — just break the density.

---

### Issue 15 — "Research" fires 5× in 12 lines
**Severity:** Medium
**Category:** Repetition / redundancy
**Location:** Section 2 (lines 47–58)
**Voice-guide rule:** §3 Sentence Architecture.
**Current text (verbatim):**
> "Do the research (or as I sometimes call it, professional stalking)."
> "the 'listen to multiple episodes' part"
> "Per the Muck Rack podcaster survey, 87% of podcasters want you to have reviewed their website, content…"
> "Skip the show if the research turns up red flags."

**Why it's an issue:** "Research / researched / research turns up / do the research" appears five times in 12 lines. The "professional stalking" joke can carry more of the load — a joke that good shouldn't have to share the load with four literal "research" mentions.

**Recommended fix:** Let "professional stalking" be the through-line. Swap two of the "research" instances for specific verbs ("scan," "audit," "sweep") and lean into the stalking joke as the connective tissue.

---

### Issue 16 — Weaker than §12.1 voice-guide anchor
**Severity:** Low
**Category:** Other editorial catches
**Location:** Line 21 (intro)
**Voice-guide rule:** §12.1 Example Opening Hook — the voice-guide-provided phrasing is "couldn't get a foot in the door of."
**Current text (verbatim):**
> "watched a few sentences land my clients on shows their competitors could not get into"

**Why it's an issue:** The voice-guide anchor paragraph in §12.1 has the stronger verb phrase "couldn't get a foot in the door of." The draft drifted to flatter phrasing despite the calibration example sitting two clicks away.

**Recommended fix:** Swap for the voice-guide phrase: "couldn't get a foot in the door of."

---

### Issue 17 — Teacher-ish admonishment
**Severity:** Low
**Category:** Meta-commentary breaks flow
**Location:** Line 41
**Voice-guide rule:** §10 Humor Register — Lexie's voice is inviting, not gatekeeping.
**Current text (verbatim):**
> "Work through these in order. Do not skip."

**Why it's an issue:** "Do not skip" reads as teacher-ish / mildly lecturing. The voice is closer to "Work through these in order" without the admonishment. It's the third narrator-interjects beat in the post (see Issue 6).

**Recommended fix:** Cut "Do not skip." The first sentence carries the full instruction.

---

### Issue 18 — "40 other pitches" is an invented number
**Severity:** Low
**Category:** Other editorial catches
**Location:** Line 118
**Voice-guide rule:** §11 DO #4 "No claim without a receipt."
**Current text (verbatim):**
> "They have 40 other pitches in the inbox today that did not."

**Why it's an issue:** "40" reads as a placeholder number rather than a sourced stat. §11 DO #4 demands a receipt on every claim. The Muck Rack data earlier in the post has real numbers — the draft is reaching for a stat and inventing one instead of reusing one.

**Recommended fix:** Either cite the Muck Rack stat on daily pitch volume, drop the specific number, or pull a named-client inbox-count anecdote if one exists in the persona file.

---

### Issue 19 — "uses" is a weak verb
**Severity:** Low
**Category:** Other editorial catches
**Location:** Line 39
**Voice-guide rule:** §12.5 example — "this is the exact workflow I've run for 15+ years" has more bite.
**Current text (verbatim):**
> "This is the exact workflow my team uses."

**Why it's an issue:** "Uses" is flat. The §12.5 voice-guide sign-off example anchors a stronger verb phrasing with a year-count receipt baked in.

**Recommended fix:** Swap to "run" and add the year-count: "This is the exact workflow I have run for 15+ years."

---

### Issue 20 — Section 4 → Section 5 transition seam
**Severity:** Low
**Category:** Transitions and flow
**Location:** Line 89
**Voice-guide rule:** Editorial: transition seams.
**Current text (verbatim):**
> "Write the pitch itself. Four parts, no filler."

**Why it's an issue:** The sentence drops in as a list-item header right after the customization example with no transitional beat. Reads like a mid-thought list item rather than a new step in a numbered sequence.

**Recommended fix:** Add a one-sentence bridge: "Now that the angle is clean, write the pitch itself. Four parts, no filler."

---

## What's Working (keep these)

- **Pre Pitch Op acronym** (lines 63–74). Named proprietary framework, clearly defined, 10 atomic items — the exact format AI Overviews and Featured Snippets love. Voice-accurate: the framework appears verbatim in the source transcripts.
- **Three-pitch customization example** (lines 83–86). "Same client, three different pitches" is genuinely the teaching moment — different audience, different angle, different headline, all traceable to one campaign. The kind of concrete receipt §6 demands.
- **"Hi Lexi" pet-peeve beat** (line 112). Verbatim from the persona file — the misspelled-name joke lands because it is specific, personal, and named. The one place in the post where the "I repeat" move earns its place.
- **Muck Rack stat drops** (line 36, line 48). Two industry stats, named source, percentages attached. §11 DO #4 compliant.
- **"One touch is a pitch. Two is a nudge. Three is stalking"** (line 105). Crisp, quotable, voice-accurate. A triple-list, but earned — the cadence matches the content (escalating frequency).
- **CTA block** (lines 159–162). "Small but mighty," "obsessed with your business," "slide into my DMs," specific service-page link to `/services/launch-pad`. All four §11 DO-list boxes ticked. "Cheers" closer present.
- **Section 5 four-part pitch anatomy** (lines 89–97). Subject / relevance line / angle / proof — four parts, example each. Scannable, numbered, voice-accurate to the Tactical Walk-Through pattern in §5.4.

---

## Systemic Patterns Worth Addressing

1. **Triple-list cadence is the dominant rhythm issue.** Sentences ending on three-item lists appear in the intro, the four-bullet anatomy block, the three-pitch example, the anatomy-of-pitch section, the FAQ section, and the recap. A single pass that rewrites every third triple-list into a different sentence shape — a short knockout, a conditional, a named receipt, or a question — would fix ~40% of the flagged cadence issues.
2. **Signature moves are stacking at 3× each.** "I repeat" + bolded-repeat fires 3×, the bolded-emphasis pattern fires 3×, the contrastive "X, not Y" closer fires 8× total. Rule of thumb for the rewrite: each signature move fires at most once per post.
3. **Hedged receipts where named-client receipts would land harder.** "Per my own HARO receipts" (line 100) and "I had a client recently" (line 105) both under-perform compared to the named-partner + percentage receipts the voice guide calibrates against in §12.3. The persona file carries named anecdotes with specifics — swap in one or two on the rewrite.
4. **Algorithmic Authorship pass was not run cleanly.** Voice guide §2.4 requires an AA pass after the voice pass. Line 34 has the condition before the main clause; lines 52–56 are noun-phrases where verb-first would extract better in an AI Overview. A 10-minute AA sweep would catch both.

---

## Signature-Move Count Check

| Move | Cap per post | Count in draft | Over cap? |
|---|:---:|:---:|:---:|
| "I repeat" pattern (literal + bolded-repeat paraphrase) | 1 | 3 | yes |
| Bolded emphasis beats (independent of "I repeat") | ≤3 typical | 5 | borderline |
| "Cheers" closer | 1 | 1 | no |
| Contrastive "X, not Y" close (FAQ + body) | ~3 typical | 8 | yes |
| Em-dash usage (§11 DON'T #11) | rare | 10 (all in acronym) | yes |
| "Slide into my DMs" | 1 | 1 | no |
| Wine / cocktail puns (§4 — Growth Mode: rare) | 0–1 | 0 | no |
receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementationStars: 322kUpdated Claude CodeCodexVerifiedInstall
requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirementsStars: 322kUpdated Claude CodeCodexVerifiedInstall
using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallbackStars: 322kUpdated Claude CodeCodexVerifiedInstall
verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions alwaysStars: 322kUpdated Claude CodeCodexVerifiedInstall