refactor: replace rigid imperatives with reasoning-based guidance

Audit all 32 SKILL.md files against Anthropic's skill-creator writing
guidance ("why > MUST" pattern). Replaces ALWAYS/NEVER/MUST/IMPORTANT
imperatives with explanations of WHY the guidance matters.

17 edits across 14 skills:
- ad-creative: character limits reasoning, CTA headline reasoning
- seo-audit: schema detection warning softened, reasoning added
- programmatic-seo: subfolder vs subdomain reasoning
- paid-ads: exclusion list reasoning
- copywriting: honesty principle reasoning
- cold-email: follow-up value reasoning
- ai-seo: freshness signal reasoning
- churn-prevention: post-cancel path reasoning
- product-marketing-context: verbatim language reasoning
- popup-cro: close button visibility reasoning
- signup-flow-cro: label visibility reasoning
- form-cro: label visibility reasoning
- revops: fallback owner reasoning
- ab-test-setup: DON'T → Avoid

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

skills/ab-test-setup/SKILL.md

@@ -180,7 +180,7 @@ We'll know this is true when [metrics].
 - Check segment quality
 - Document external factors
 
-**DON'T:**
+**Avoid:**
 - Peek at results and stop early
 - Make changes to variants
 - Add traffic from new sources

skills/ad-creative/SKILL.md

@@ -64,7 +64,7 @@ Pull performance data → Identify winning patterns → Generate new variations
 
 ## Platform Specs
 
-**Always enforce these limits.** Never deliver creative that exceeds platform character limits.
+Platforms reject or truncate creative that exceeds these limits, so verify every piece of copy fits before delivering.
 
 ### Google Ads (Responsive Search Ads)
 
@@ -320,7 +320,7 @@ For large-scale creative production (Anthropic's growth team generates 100+ vari
 - **Writing headlines that only work together** — RSA headlines get combined randomly
 - **Ignoring character limits** — Platforms truncate without warning
 - **All variations sound the same** — Vary angles, not just word choice
-- **No CTA headlines** — Always include action-oriented headlines
+- **No CTA headlines** — RSAs need action-oriented headlines to drive clicks; include at least 2-3
 - **Generic descriptions** — "Learn more about our solution" wastes the slot
 - **Iterating without data** — Gut feelings are less reliable than metrics
 - **Testing too many things at once** — Change one variable per test cycle

skills/ai-seo/SKILL.md

@@ -353,7 +353,7 @@ Monthly manual check:
 - **Ignoring AI search entirely** — ~45% of Google searches now show AI Overviews, and ChatGPT/Perplexity are growing fast
 - **Treating AI SEO as separate from SEO** — Good traditional SEO is the foundation; AI SEO adds structure and authority on top
 - **Writing for AI, not humans** — If content reads like it was written to game an algorithm, it won't get cited or convert
-- **No freshness signals** — Undated content loses to dated content. Always show when content was last updated
+- **No freshness signals** — Undated content loses to dated content because AI systems weight recency heavily. Show when content was last updated
 - **Gating all content** — AI can't access gated content. Keep your most authoritative content open
 - **Ignoring third-party presence** — You may get more AI citations from a Wikipedia mention than from your own blog
 - **No structured data** — Schema markup gives AI systems structured context about your content

skills/churn-prevention/SKILL.md

@@ -375,7 +375,7 @@ Test one variable at a time:
 - **Guilt-trip copy** — "Are you sure you want to abandon us?" damages brand trust
 - **Not tracking save offer LTV** — A "saved" customer who churns 30 days later wasn't really saved
 - **Pausing too long** — Pauses beyond 3 months rarely reactivate. Set limits.
-- **No post-cancel path** — Always make reactivation easy and trigger win-back emails
+- **No post-cancel path** — Make reactivation easy and trigger win-back emails, because some churned users will want to come back
 
 ---
 

skills/cold-email/SKILL.md

@@ -100,7 +100,7 @@ See [subject-lines.md](references/subject-lines.md) for the full data.
 
 ## Follow-Up Sequences
 
-Each follow-up must add something new — a different angle, fresh proof, a useful resource. Never "just checking in."
+Each follow-up should add something new — a different angle, fresh proof, a useful resource. "Just checking in" gives the reader no reason to respond.
 
 - 3-5 total emails, increasing gaps between them
 - Each email should stand alone (they may not have read the previous ones)

skills/copywriting/SKILL.md

@@ -67,7 +67,7 @@ Each section should advance one argument. Build a logical flow down the page.
 3. **Active over passive** — "We generate reports" not "Reports are generated"
 4. **Confident over qualified** — Remove "almost," "very," "really"
 5. **Show over tell** — Describe the outcome instead of using adverbs
-6. **Honest over sensational** — Never fabricate statistics or testimonials
+6. **Honest over sensational** — Fabricated statistics or testimonials erode trust and create legal liability
 
 ### Quick Quality Check
 

skills/form-cro/SKILL.md

@@ -120,7 +120,7 @@ For each field, ask:
 4. Logical grouping if many fields
 
 ### Labels and Placeholders
-- Labels: Always visible (not just placeholder)
+- Labels: Keep visible (not just placeholder) — placeholders disappear when typing, leaving users unsure what they're filling in
 - Placeholders: Examples, not labels
 - Help text: Only when genuinely helpful
 

skills/paid-ads/SKILL.md

@@ -122,7 +122,7 @@ LI_LeadGen_CMOs-SaaS_Whitepaper_Mar24
 
 - **Lookalikes**: Base on best customers (by LTV), not all customers
 - **Retargeting**: Segment by funnel stage (visitors vs. cart abandoners)
-- **Exclusions**: Always exclude existing customers and recent converters
+- **Exclusions**: Exclude existing customers and recent converters — showing ads to people who already bought wastes spend
 
 **For detailed targeting strategies by platform**: See [references/audience-targeting.md](references/audience-targeting.md)
 

skills/popup-cro/SKILL.md

@@ -182,7 +182,7 @@ Before providing recommendations, understand:
 - Leave space to close (visible X, click outside)
 
 ### Close Button
-- Always visible (top right is convention)
+- Keep visible (top right is convention) — users who can't find the close button will bounce entirely
 - Large enough to tap on mobile
 - "No thanks" text link as alternative
 - Click outside to close

skills/product-marketing-context/SKILL.md

@@ -47,7 +47,7 @@ For each section:
 3. Confirm accuracy
 4. Move to the next
 
-**Important:** Push for verbatim customer language. Exact phrases are more valuable than polished descriptions.
+Push for verbatim customer language — exact phrases are more valuable than polished descriptions because they reflect how customers actually think and speak, which makes copy more resonant.
 
 ---
 

skills/programmatic-seo/SKILL.md

@@ -49,7 +49,7 @@ Hierarchy of data defensibility:
 5. Public (anyone can use—weakest)
 
 ### 3. Clean URL Structure
-**Always use subfolders, not subdomains**:
+**Use subfolders, not subdomains** — subfolders consolidate domain authority while subdomains split it:
 - Good: `yoursite.com/templates/resume/`
 - Bad: `templates.yoursite.com/resume/`
 

skills/revops/SKILL.md

@@ -133,7 +133,7 @@ Define response times and document them:
 ### Routing Rules Essentials
 
 - Route to the **most specific match** first, then fall back to general
-- Always include a **fallback owner** — no lead should go unassigned
+- Include a **fallback owner** — unassigned leads go cold fast and waste pipeline
 - Round-robin should account for **rep capacity and availability** (PTO, quota attainment)
 - Log every routing decision for audit and optimization
 

skills/seo-audit/SKILL.md

@@ -35,7 +35,7 @@ Before auditing, understand:
 
 ## Audit Framework
 
-### ⚠️ Important: Schema Markup Detection Limitation
+### Schema Markup Detection Limitation
 
 **`web_fetch` and `curl` cannot reliably detect structured data / schema markup.**
 
@@ -46,7 +46,7 @@ Many CMS plugins (AIOSEO, Yoast, RankMath) inject JSON-LD via client-side JavaSc
 2. **Google Rich Results Test** — https://search.google.com/test/rich-results
 3. **Screaming Frog export** — if the client provides one, use it (SF renders JavaScript)
 
-**Never report "no schema found" based solely on `web_fetch` or `curl`.** This has led to false audit findings in production.
+Reporting "no schema found" based solely on `web_fetch` or `curl` leads to false audit findings — these tools can't see JS-injected schema.
 
 ### Priority Order
 1. **Crawlability & Indexation** (can Google find and index it?)
@@ -382,7 +382,7 @@ Same format as above
 - Mobile-Friendly Test
 - Schema Validator
 
-> **Note on schema detection:** `web_fetch` strips `<script>` tags (including JSON-LD) and cannot detect JS-injected schema. Always use the browser tool, Rich Results Test, or Screaming Frog for schema checks. See the warning at the top of the Audit Framework section.
+> **Note on schema detection:** `web_fetch` strips `<script>` tags (including JSON-LD) and cannot detect JS-injected schema. Use the browser tool, Rich Results Test, or Screaming Frog instead — they render JavaScript and capture dynamically-injected markup. See the Schema Markup Detection Limitation section above.
 
 **Paid Tools** (if available)
 - Screaming Frog

skills/signup-flow-cro/SKILL.md

@@ -161,7 +161,7 @@ Every field reduces conversion. For each field, ask:
 
 ### Microcopy
 - Placeholder text: Use for examples, not labels
-- Labels: Always visible (not just placeholders)
+- Labels: Keep visible (not just placeholders) — placeholders disappear when typing, leaving users unsure what they're filling in
 - Help text: Only when needed, placed close to field
 
 ---