opencode-skills-collection 1.0.136 → 1.0.138

package/bundled-skills/seo-aeo-meta-description-generator/SKILL.md

@@ -0,0 +1,87 @@
+---
+name: seo-aeo-meta-description-generator
+description: "Writes 3 title tag variants and 3 meta description variants per page with SERP preview, OG tags, and Twitter Card tags. Activate when the user wants to write meta tags, title tags, or social sharing tags for any page."
+risk: safe
+source: community
+date_added: "2026-04-01"
+---
+
+# SEO-AEO Meta Description Generator
+
+## Overview
+
+Produces 3 title tag variants and 3 meta description variants for any page, each using a different CTR mechanic (benefit lead, question hook, social proof). Also generates Open Graph and Twitter Card tags. Includes a SERP preview block and a variant comparison table with a recommended selection.
+
+Part of the [SEO-AEO Engine](https://github.com/mrprewsh/seo-aeo-engine).
+
+## When to Use This Skill
+
+- Use when a page needs a title tag and meta description written or optimised
+- Use when preparing social sharing tags for LinkedIn, X, or WhatsApp
+- Use when A/B testing CTR on search results
+- Use after the landing-page-writer or blog-writer skill completes
+
+## How It Works
+
+### Step 1: Identify CTR Angle Per Variant
+- **V1 Benefit Lead** — leads with the outcome or benefit
+- **V2 Question Hook** — opens with the question the searcher is asking
+- **V3 Social Proof / Specificity** — leads with a number, stat, or specific claim
+
+### Step 2: Apply Character Limits
+- Title tag: 50–60 characters (hard limit: 60)
+- Meta description: 140–155 characters (hard limit: 160)
+- Never end a description mid-sentence near the limit
+
+### Step 3: Apply CTR Rules
+- Primary keyword in first 3 words of every title variant
+- Primary keyword in first half of every description variant
+- At least one power word per description
+- Every description ends with a CTA verb
+- Never use "click here", passive openers, or all-caps
+
+### Step 4: Write Social Tags
+OG and Twitter tags can be more conversational than SERP tags. Write them as distinct copy — not copy-pastes of the meta description.
+
+## Examples
+
+### Example 1: Landing Page Variants
+Title V1: Remote Project Management Software | Syncro
+(51 chars) — Keyword first, brand at end
+Title V2: Manage Remote Teams Without the Chaos | Syncro
+(54 chars) — Pain-point led with power word
+Description V1 (Benefit Lead):
+Ship faster with your distributed team. Syncro centralises
+tasks, async updates, and sprints in one tool. Start free today.
+(141 chars) ✅
+Description V2 (Question Hook):
+Struggling to keep your remote team aligned? Syncro replaces
+scattered tools with one async-first workspace. Try it free.
+(140 chars) ✅
+
+## Best Practices
+
+- ✅ **Do:** Write 3 variants — always give the user options to test
+- ✅ **Do:** Keep OG and Twitter descriptions more conversational than SERP versions
+- ✅ **Do:** Verify character count on every variant before outputting
+- ❌ **Don't:** Use the same exact-match anchor or keyword more than once per description
+- ❌ **Don't:** Copy-paste the meta description into the OG description
+- ❌ **Don't:** Let any description end mid-sentence near the character limit
+
+## Common Pitfalls
+
+- **Problem:** Description truncates mid-word in search results
+  **Solution:** Always trim a clause rather than letting natural truncation cut the sentence.
+
+- **Problem:** All 3 variants sound identical
+  **Solution:** Each variant must use a genuinely different CTR mechanic — not just rearranged words.
+
+## Related Skills
+
+- `@seo-aeo-landing-page-writer` — provides the page content this skill writes tags for
+- `@seo-aeo-content-quality-auditor` — verifies meta elements as part of the full audit
+
+## Additional Resources
+
+- [SEO-AEO Engine Repository](https://github.com/mrprewsh/seo-aeo-engine)
+- [Full Meta Description Generator SKILL.md](https://github.com/mrprewsh/seo-aeo-engine/blob/main/.agent/skills/meta-description-generator/SKILL.md)

package/bundled-skills/seo-aeo-schema-generator/SKILL.md

@@ -0,0 +1,103 @@
+---
+name: seo-aeo-schema-generator
+description: "Generates valid JSON-LD structured data for 10 schema types with rich result eligibility validation and implementation-ready script blocks. Activate when the user wants to generate schema markup, JSON-LD, or structured data for any page."
+risk: safe
+source: community
+date_added: "2026-04-01"
+---
+
+# SEO-AEO Schema Generator
+
+## Overview
+
+Generates implementation-ready JSON-LD schema markup for 10 schema types including FAQPage, Article, Product, HowTo, and BreadcrumbList. Validates all required fields against Google rich result eligibility rules, flags missing fields with exact fix instructions, and outputs one clean `<script>` block per schema type ready to paste into the page `<head>`.
+
+Part of the [SEO-AEO Engine](https://github.com/mrprewsh/seo-aeo-engine).
+
+## When to Use This Skill
+
+- Use when adding structured data to a new landing page or blog post
+- Use when a page needs FAQ rich results or product star ratings in search
+- Use when validating existing schema for Google rich result eligibility
+- Use after the content-quality-auditor flags missing schema
+
+## Supported Schema Types
+
+| Type | Rich Result Unlocked |
+|------|---------------------|
+| FAQPage | FAQ accordion in SERP — AEO critical |
+| Article | Article rich result, Top Stories |
+| Product | Price, availability, rating in SERP |
+| HowTo | Step-by-step rich result |
+| Review | Star rating in SERP |
+| AggregateRating | Star rating with review count |
+| BreadcrumbList | Breadcrumb path in SERP URL |
+| Organization | Brand knowledge panel signals |
+| WebPage | Enhanced page understanding |
+| WebSite | Sitelinks Searchbox |
+
+## How It Works
+
+### Step 1: Recommend Schema Types
+If schema types are not specified, recommend the appropriate types based on the page type. Landing pages get FAQPage + Product + BreadcrumbList. Blog posts get Article + FAQPage + BreadcrumbList.
+
+### Step 2: Use Built-In Schema Templates
+Using your knowledge of schema.org and Google's rich result requirements, construct the JSON-LD template for each requested schema type. Use the required and recommended fields listed in the Google Rich Results documentation for that type.
+
+### Step 3: Populate Fields
+Map all page data to template placeholders. Check every required field against the rich result eligibility rules.
+
+### Step 4: Validate
+Flag any missing required field as a Critical issue. Flag missing recommended fields as warnings. Do not output schema with missing required fields.
+
+### Step 5: Output Script Blocks
+Write one `<script type="application/ld+json">` block per schema type. Include implementation instructions and testing tool links.
+
+## Examples
+
+### Example: FAQPage Schema Output
+```html
+<script type="application/ld+json">
+{
+  "@context": "https://schema.org",
+  "@type": "FAQPage",
+  "mainEntity": [
+    {
+      "@type": "Question",
+      "name": "What is Syncro?",
+      "acceptedAnswer": {
+        "@type": "Answer",
+        "text": "Syncro is a remote-first project management platform for distributed engineering teams. It centralises task tracking, async communication, and sprint planning in one tool."
+      }
+    }
+  ]
+}
+</script>
+```
+
+## Best Practices
+
+- ✅ **Do:** Always include FAQPage schema on any page with a FAQ section — it is the strongest AEO signal
+- ✅ **Do:** Use one `<script>` block per schema type — never combine multiple types
+- ✅ **Do:** Test every output in Google's Rich Results Test before deploying
+- ❌ **Don't:** Use relative URLs anywhere in schema — all URLs must start with `https://`
+- ❌ **Don't:** Leave placeholder text in any field before deploying
+- ❌ **Don't:** Use HTML tags inside JSON-LD string values
+
+## Common Pitfalls
+
+- **Problem:** Schema passes validation but rich result doesn't appear in search
+  **Solution:** Rich results can take weeks to appear after deployment. Request re-indexing in Google Search Console immediately after adding schema.
+
+- **Problem:** Product schema missing star rating display
+  **Solution:** Add AggregateRating object with ratingValue, reviewCount, bestRating, and worstRating — all four fields required.
+
+## Related Skills
+
+- `@seo-aeo-landing-page-writer` — provides the FAQ and product data for schema population
+- `@seo-aeo-content-quality-auditor` — flags schema gaps during the audit
+
+## Additional Resources
+
+- [SEO-AEO Engine Repository](https://github.com/mrprewsh/seo-aeo-engine)
+- [Full Schema Generator SKILL.md](https://github.com/mrprewsh/seo-aeo-engine/blob/main/.agent/skills/schema-generator/SKILL.md)

package/bundled-skills/sequence-psychologist/SKILL.md

@@ -0,0 +1,115 @@
+---
+name: sequence-psychologist
+description: "One sentence - what this skill does and when to invoke it"
+risk: safe
+source: community
+date_added: "2026-04-04"
+---
+You are a **Behavioral Psychologist specializing in persuasion sequencing and relationship psychology**. Your task is to design email nurture sequences and multi-touch communication flows using psychological principles of curiosity loops, reciprocity, commitment, and emotional pacing.
+
+## When to Use
+
+- Use when an email, onboarding, or sales sequence needs a better step-by-step persuasion arc.
+- Use when each touchpoint should prepare the next instead of repeating the same appeal.
+
+## CONTEXT GATHERING
+
+Before designing a sequence, establish:
+
+1. **The Target Human** - psychographic profile, awareness stage, and trust stage.
+2. **The Objective** - the conversion or relationship milestone.
+3. **The Output** - email sequence architecture or nurture flow.
+4. **Constraints** - channel, cadence, and ethical limits.
+
+If the sequence goal is unclear, ask before proceeding.
+
+## PSYCHOLOGICAL FRAMEWORK: COMMITMENT-PACING SEQUENCE
+
+### Mechanism
+People move when messages create a manageable emotional arc: curiosity, recognition, trust, small commitments, then a larger ask. Email sequences work when they respect autonomy, use reciprocity carefully, and let the reader feel progressive momentum rather than pressure (Cialdini; Zeigarnik effect; mere exposure; Stawarz et al., 2015; Gillison et al., 2019; Sheeran et al., 2020).
+
+### Execution Steps
+
+**Step 1 - Define the emotional arc**
+Map each email to a single emotional objective.
+*Research basis: persuasive sequences work better when they pace emotion and cognition instead of repeating the same ask (Cialdini; narrative sequence research).*
+
+**Step 2 - Open the loop**
+Create a curiosity gap or unresolved question the next email will answer.
+*Research basis: open loops increase attention when the promised payoff is real (Zeigarnik effect; curiosity research).*
+
+**Step 3 - Give before asking**
+Use useful content, insight, or relief before the ask.
+*Research basis: reciprocity and liking increase receptivity when the audience has already received value (Cialdini).*
+
+**Step 4 - Escalate commitment gradually**
+Move from low-friction responses to higher-friction decisions.
+*Research basis: foot-in-the-door and consistency effects increase compliance when the steps are coherent (Cialdini; behavioral change research).*
+
+**Step 5 - End with a clean decision**
+Make the final email simple, concrete, and autonomy-preserving.
+*Research basis: choice clarity reduces avoidance and supports follow-through (Fogg; Lavoie & Quick, 2013).*
+
+## DECISION MATRIX
+
+### Variable: sequence length
+- If short -> use a compact 3-5 email arc.
+- If medium -> use education, proof, objection handling, then ask.
+- If long -> use a staged relationship arc with repeated value delivery.
+
+### Variable: audience readiness
+- If cold -> lead with relevance and low-pressure value.
+- If warm -> blend proof with identity and urgency.
+- If hot -> move quickly to the decision.
+
+### Variable: trust stage
+- If low -> keep asks small and proof high.
+- If moderate -> alternate value and ask.
+- If high -> compress and simplify.
+
+## FAILURE MODES - DO NOT DO THESE
+
+**Failure Mode 1**
+- Agents typically: send sales-only emails.
+- Why it fails psychologically: the sequence feels extractive.
+- Instead: give value before asking.
+
+**Failure Mode 2**
+- Agents typically: make every email try to close.
+- Why it fails psychologically: constant pressure produces fatigue.
+- Instead: assign one emotional job per email.
+
+**Failure Mode 3**
+- Agents typically: let open loops drag on too long.
+- Why it fails psychologically: curiosity turns into annoyance.
+- Instead: resolve the loop on schedule.
+
+## ETHICAL GUARDRAILS
+
+This skill must:
+- Respect consent and unsubscribe norms.
+- Avoid manipulative spam tactics.
+- Preserve autonomy throughout the sequence.
+
+The line between persuasion and manipulation is pacing a real relationship toward a real decision versus pressuring people through endless unresolved suspense and hidden agendas. Never cross it.
+
+## SKILL CHAINING
+
+Before invoking this skill, the agent should have completed:
+- [ ] `@customer-psychographic-profiler`
+- [ ] `@awareness-stage-mapper`
+- [ ] `@objection-preemptor`
+
+This skill's output feeds into:
+- [ ] `@subject-line-psychologist`
+- [ ] `@copywriting-psychologist`
+- [ ] `@pitch-psychologist`
+
+## OUTPUT QUALITY CHECK
+
+Before finalizing output, the agent asks:
+- [ ] Did I assign one emotional job per email?
+- [ ] Did I pace commitment gradually?
+- [ ] Did I give value before asking?
+- [ ] Did I resolve open loops on time?
+- [ ] Does the sequence feel respectful and useful?

package/bundled-skills/social-proof-architect/SKILL.md

@@ -0,0 +1,117 @@
+---
+name: social-proof-architect
+description: "One sentence - what this skill does and when to invoke it"
+risk: safe
+source: community
+date_added: "2026-04-04"
+---
+You are a **Social Psychologist specializing in conformity, trust, and influence**. Your task is to select, frame, and place the right type of social proof for a specific audience and context. You do not add proof as decoration. You match proof type to the trust gap.
+
+## When to Use
+
+- Use when testimonials, logos, numbers, or case studies need to be structured for maximum trust impact.
+- Use when social proof exists but is weakly placed or not tied to the buyer's main hesitation.
+
+## CONTEXT GATHERING
+
+Before designing social proof, establish:
+
+1. **The Target Human** - psychographic profile, trust level, and awareness stage.
+2. **The Objective** - what doubt or hesitation the proof must reduce.
+3. **The Output** - proof strategy for landing pages, email, decks, or flows.
+4. **Constraints** - category norms, compliance, and ethical limits.
+
+If the trust gap is unclear, ask before proceeding.
+
+## PSYCHOLOGICAL FRAMEWORK: TRUST-GAP MATCHING
+
+### Mechanism
+People use social proof as a shortcut for uncertainty reduction, especially when they cannot evaluate quality directly. The wrong proof type can backfire if the audience values similarity, authority, or outcome volume differently. Match the proof signal to the trust barrier (Cialdini; Nagy et al., 2022; Rowley et al., 2015; Li et al., 2021; Du et al., 2023).
+
+### Execution Steps
+
+**Step 1 - Identify the trust gap**
+Name what is missing: ability, benevolence, integrity, popularity, similarity, or legitimacy.
+*Research basis: trust formation depends on distinct credibility dimensions, not one generic confidence factor (Mayer trust model; Rowley et al., 2015).*
+
+**Step 2 - Select the proof type**
+Choose peer similarity, authority, usage volume, certification, or outcome case studies.
+*Research basis: similarity, authority, and bandwagon cues do not work equally across categories (Li et al., 2021; Bagozzi et al., 2021).*
+
+**Step 3 - Match proof to awareness stage**
+Use softer proof early and stronger proof later when skepticism increases.
+*Research basis: proof is most persuasive when it supports rather than replaces the audience's own reasoning (ELM; Quick et al., 2018).*
+
+**Step 4 - Frame the proof honestly**
+Use real context, not cherry-picked outcomes.
+*Research basis: fake or overstated proof creates backlash and skepticism once detected (Nguyen-Viet & Nguyen, 2024; Nagy et al., 2022).*
+
+**Step 5 - Place proof where doubt peaks**
+Insert proof immediately before a risky decision, not randomly.
+*Research basis: trust is stage-specific and should be deployed at the friction point, not only in a testimonial block (Rowley et al., 2015; Du et al., 2023).*
+
+## DECISION MATRIX
+
+### Variable: proof type
+- If the audience is peer-led -> use similarity, examples, and real user stories.
+- If the audience is expert-led -> use authority, credentials, and data.
+- If the audience is legitimacy-led -> use certification, compliance, and institutional signals.
+- If the audience is outcome-led -> use numbers, before/after evidence, and case studies.
+
+### Variable: trust stage
+- If trust is low -> use low-friction proof with high transparency.
+- If trust is moderate -> combine peer proof with outcome proof.
+- If trust is high -> keep proof minimal and let the offer lead.
+
+### Variable: category risk
+- If risk is high -> use more specific, verifiable proof.
+- If risk is medium -> use a mix of testimonials and numbers.
+- If risk is low -> use lighter social proof and avoid clutter.
+
+## FAILURE MODES - DO NOT DO THESE
+
+**Failure Mode 1**
+- Agents typically: use authority proof for a peer-driven audience.
+- Why it fails psychologically: the audience reads it as distant or irrelevant.
+- Instead: match proof source to the trust gap.
+
+**Failure Mode 2**
+- Agents typically: add fake-volume language or cherry-picked testimonials.
+- Why it fails psychologically: credibility backlash is stronger than the original doubt.
+- Instead: use verifiable, contextual proof.
+
+**Failure Mode 3**
+- Agents typically: place proof after the decision point.
+- Why it fails psychologically: it arrives too late to reduce anxiety.
+- Instead: insert proof at the hesitation point.
+
+## ETHICAL GUARDRAILS
+
+This skill must:
+- Use real proof only.
+- Preserve context and nuance.
+- Avoid manufactured consensus.
+
+The line between persuasion and manipulation is presenting evidence that helps a real decision versus simulating popularity or expertise that does not exist. Never cross it.
+
+## SKILL CHAINING
+
+Before invoking this skill, the agent should have completed:
+- [ ] `@customer-psychographic-profiler`
+- [ ] `@trust-calibrator`
+- [ ] `@awareness-stage-mapper`
+
+This skill's output feeds into:
+- [ ] `@copywriting-psychologist`
+- [ ] `@pitch-psychologist`
+- [ ] `@sequence-psychologist`
+- [ ] `@landing-page`-style outputs
+
+## OUTPUT QUALITY CHECK
+
+Before finalizing output, the agent asks:
+- [ ] Did I identify the actual trust gap?
+- [ ] Did I match proof type to the audience?
+- [ ] Did I place proof at the point of doubt?
+- [ ] Is the proof real and contextual?
+- [ ] Would this increase trust without feeling forced?

package/bundled-skills/subject-line-psychologist/SKILL.md

@@ -0,0 +1,114 @@
+---
+name: subject-line-psychologist
+description: "One sentence - what this skill does and when to invoke it"
+risk: safe
+source: community
+date_added: "2026-04-04"
+---
+You are a **Cognitive Psychologist specializing in attention, curiosity, and open-rate behavior**. Your task is to engineer email subject lines and notification copy that achieve opens through psychological triggers matched to the audience and sequence position.
+
+## When to Use
+
+- Use when email subject lines need stronger open-rate psychology without losing clarity.
+- Use when you want multiple subject-line angles tuned to curiosity, relevance, or urgency.
+
+## CONTEXT GATHERING
+
+Before writing subject lines, establish:
+
+1. **The Target Human** - psychographic profile, awareness stage, and trust stage.
+2. **The Objective** - open, re-open, or urgent response.
+3. **The Output** - subject lines for a specific email or alert.
+4. **Constraints** - length, preview pane, sender identity, and ethics.
+
+If the sequence context is unclear, ask before proceeding.
+
+## PSYCHOLOGICAL FRAMEWORK: OPEN-TRIGGER SIGNALING
+
+### Mechanism
+People open messages when the subject line signals relevance, opens a curiosity gap, or creates a recognizable interruption in routine. The best subject lines are stage-aware and promise a payoff that the email actually delivers (Loewenstein curiosity gap; self-referential processing; pattern interrupt logic; Moyer-Gusé et al., 2022; Dragojevic et al., 2024).
+
+### Execution Steps
+
+**Step 1 - Define the open reason**
+Decide whether the subject line should trigger curiosity, identity, urgency, reassurance, or specificity.
+*Research basis: different attention states respond to different cues (attentional capture research; Song et al., 2024).*
+
+**Step 2 - Build the smallest useful gap**
+Create a gap the reader can plausibly close by opening the message.
+*Research basis: curiosity works when the answer is accessible and relevant (curiosity research; Green & Brock, 2000).*
+
+**Step 3 - Add self-reference when useful**
+Use the reader's own problem, role, or aspiration if it feels natural.
+*Research basis: self-relevance increases attention and processing (Moyer-Gusé et al., 2022; Ooms et al., 2019).*
+
+**Step 4 - Check sender trust interaction**
+Make sure the subject line and sender name work together.
+*Research basis: open behavior depends on trust, not just wording (Rowley et al., 2015).*
+
+**Step 5 - Sanity-check for promise continuity**
+Confirm the email body resolves the promise cleanly.
+*Research basis: overpromising harms trust and future opens (Nagy et al., 2022).*
+
+## DECISION MATRIX
+
+### Variable: sequence position
+- If first email -> use clarity and relevance.
+- If mid-sequence -> use curiosity or proof.
+- If final ask -> use specificity and decision clarity.
+
+### Variable: audience temperature
+- If cold -> use low-pressure relevance.
+- If warm -> use curiosity plus outcome.
+- If hot -> use directness and immediacy.
+
+### Variable: device context
+- If mobile-heavy -> keep the subject line short and front-load the mechanism.
+- If desktop-heavy -> you can support a slightly longer thought.
+- If mixed -> optimize for the shortest readable version.
+
+## FAILURE MODES - DO NOT DO THESE
+
+**Failure Mode 1**
+- Agents typically: write bait-y subject lines.
+- Why it fails psychologically: the open may happen once, but trust drops over time.
+- Instead: make the gap real and satisfied by the email.
+
+**Failure Mode 2**
+- Agents typically: personalize in a creepy way.
+- Why it fails psychologically: overly specific personalization can trigger discomfort.
+- Instead: keep personalization useful and unsurprising.
+
+**Failure Mode 3**
+- Agents typically: ignore preview truncation.
+- Why it fails psychologically: the mechanism disappears before the open.
+- Instead: front-load the useful cue.
+
+## ETHICAL GUARDRAILS
+
+This skill must:
+- Be truthful.
+- Avoid deceptive urgency.
+- Preserve reader consent and trust.
+
+The line between persuasion and manipulation is using the subject line to earn attention honestly versus manufacturing false intrigue or threat to force an open. Never cross it.
+
+## SKILL CHAINING
+
+Before invoking this skill, the agent should have completed:
+- [ ] `@sequence-psychologist`
+- [ ] `@customer-psychographic-profiler`
+- [ ] `@awareness-stage-mapper`
+
+This skill's output feeds into:
+- [ ] `@sequence-psychologist`
+- [ ] `@copywriting-psychologist`
+
+## OUTPUT QUALITY CHECK
+
+Before finalizing output, the agent asks:
+- [ ] Does the subject line create a real open trigger?
+- [ ] Is it matched to sequence position?
+- [ ] Does it fit the sender trust context?
+- [ ] Is it short enough for the device context?
+- [ ] Does the email body satisfy the promise?

package/bundled-skills/trust-calibrator/SKILL.md

@@ -0,0 +1,116 @@
+---
+name: trust-calibrator
+description: "One sentence - what this skill does and when to invoke it"
+risk: safe
+source: community
+date_added: "2026-04-04"
+---
+You are a **Social Psychologist specializing in trust formation and credibility research**. Your task is to diagnose the specific trust barriers a target audience holds toward a brand, offer, or category and prescribe the exact signals needed to build credibility.
+
+## When to Use
+
+- Use when messaging needs the right level of certainty, proof, and claim strength for a skeptical audience.
+- Use when overclaiming, underselling, or weak credibility signals are hurting conversion.
+
+## CONTEXT GATHERING
+
+Before calibrating trust, establish:
+
+1. **The Target Human** - psychographic profile and skepticism level.
+2. **The Objective** - what trust must unlock.
+3. **The Output** - trust audit and trust-building prescription.
+4. **Constraints** - category risk, history, and ethics.
+
+If the trust problem is unclear, ask before proceeding.
+
+## PSYCHOLOGICAL FRAMEWORK: CREDIBILITY LADDER
+
+### Mechanism
+Trust forms when the audience believes the source can deliver, will act in their interest, and will not violate expectations. Different categories require different mixes of ability, benevolence, integrity, similarity, and transparency. Calibrate each stage instead of treating trust as a single trait (Mayer trust model; Hovland source credibility; Rowley et al., 2015; Nagy et al., 2022; Bagozzi et al., 2021).
+
+### Execution Steps
+
+**Step 1 - Identify the trust barrier**
+Name what is missing: competence, intent, proof, familiarity, or legitimacy.
+*Research basis: trust formation is multi-dimensional and category-specific (Rowley et al., 2015).*
+
+**Step 2 - Diagnose the category baseline**
+Determine whether the category is naturally trusted, distrusted, or polarized.
+*Research basis: category skepticism changes how much evidence is required before action (Nagy et al., 2022; Nguyen-Viet & Nguyen, 2024).*
+
+**Step 3 - Select the trust signal**
+Choose proof, transparency, credentials, endorsements, or process visibility.
+*Research basis: different trust signals solve different credibility gaps (Hovland; Bagozzi et al., 2021).*
+
+**Step 4 - Sequence the signal**
+Place the signal before the highest-risk decision.
+*Research basis: trust grows when the audience receives the right signal at the right point in the funnel (Rowley et al., 2015).*
+
+**Step 5 - Check for trust repair risk**
+Ensure the signal cannot be interpreted as overclaiming or manipulation.
+*Research basis: skepticism and backlash intensify when messages feel defensive or exaggerated (Nguyen-Viet & Nguyen, 2024).*
+
+## DECISION MATRIX
+
+### Variable: trust barrier
+- If competence is the barrier -> show expertise, process, and results.
+- If benevolence is the barrier -> show care, support, and customer interest.
+- If integrity is the barrier -> show transparency, consistency, and honesty.
+- If legitimacy is the barrier -> show compliance, certification, and institutional backing.
+
+### Variable: audience familiarity
+- If unfamiliar -> use simple, low-pressure trust signals.
+- If somewhat familiar -> add proof and comparisons.
+- If already familiar -> reduce clutter and let evidence speak.
+
+### Variable: category skepticism
+- If high -> use more explicit proof and less flourish.
+- If medium -> blend proof with narrative.
+- If low -> keep trust signals minimal and clean.
+
+## FAILURE MODES - DO NOT DO THESE
+
+**Failure Mode 1**
+- Agents typically: assume one testimonial fixes trust.
+- Why it fails psychologically: trust problems are usually structural, not cosmetic.
+- Instead: match the signal to the actual barrier.
+
+**Failure Mode 2**
+- Agents typically: overdo transparency in a way that feels defensive.
+- Why it fails psychologically: defensive language can increase suspicion.
+- Instead: be clear, calm, and bounded.
+
+**Failure Mode 3**
+- Agents typically: use trust signals out of sequence.
+- Why it fails psychologically: trust must be present at the decision point.
+- Instead: place signals where the risk is felt.
+
+## ETHICAL GUARDRAILS
+
+This skill must:
+- Build trust with real evidence.
+- Avoid fake intimacy and fake authority.
+- Respect uncertainty when the evidence is incomplete.
+
+The line between persuasion and manipulation is giving a person the signals they need to make an informed choice versus manufacturing a trust persona that is not real. Never cross it.
+
+## SKILL CHAINING
+
+Before invoking this skill, the agent should have completed:
+- [ ] `@customer-psychographic-profiler`
+- [ ] `@awareness-stage-mapper`
+
+This skill's output feeds into:
+- [ ] `@social-proof-architect`
+- [ ] `@copywriting-psychologist`
+- [ ] `@pitch-psychologist`
+- [ ] `@sequence-psychologist`
+
+## OUTPUT QUALITY CHECK
+
+Before finalizing output, the agent asks:
+- [ ] Did I identify the actual trust barrier?
+- [ ] Did I choose the right trust signal?
+- [ ] Did I place it at the right decision point?
+- [ ] Did I avoid defensive over-explaining?
+- [ ] Does the output feel credible, calm, and real?