PrePilot Lite

View saved version →

---
name: analytics-measurement
description: Use when the user needs KPI design, tracking plans, events, UTMs, attribution, dashboards, campaign measurement, experiment readbacks, anomaly analysis, CAC/LTV/ROAS logic, or a way to connect marketing activity to business outcomes.
---

# Analytics and Measurement

## Use it for

- Measurement planning
- Tracking-plan design
- Attribution interpretation
- KPI trees and dashboards
- Performance readbacks and anomaly diagnosis

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Track for decisions, not because data can be collected.
- Define the business question before the metric.
- Keep platform, analytics, CRM, and finance measures distinct when they answer different questions.
- Do not reconcile conflicting numbers by averaging them.

## Workflow

1. Define the decision, business outcome, funnel model, time window, and source of truth.
2. Build a KPI tree from business outcome to leading indicators and diagnostics.
3. Specify events, properties, UTM rules, identities, and conversion definitions when tracking is in scope.
4. Identify attribution limits, missing data, and known biases.
5. Create readback format with baseline, candidate, segment, caveats, and verdict.
6. Set review cadence and anomaly thresholds appropriate to data volume.

## Output contract

- KPI tree
- Tracking or measurement plan
- Attribution caveats
- Dashboard/readback specification
- Decision rules

## Quality gate

- Every metric has a decision owner or use
- Definitions are unambiguous
- Data-source conflicts are surfaced
- Vanity metrics are not treated as business outcomes

## North Star and input metrics

When the user asks for a North Star Metric, choose one customer-value outcome that can act as a leading indicator of durable business value. Do not use a bundle of metrics and do not confuse the North Star with an OKR or a revenue target. Add a small set of input metrics that teams can influence and explain the causal hypothesis connecting each input to the North Star. Validate the choice against the business model, user value, controllability, measurability, and leading-indicator quality.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: blog-engine
description: Use when the user wants blog strategy, topic selection, briefs, outlines, article writing, editing, fact-checking, content refreshes, internal linking, SEO alignment, GEO readiness, or repurposing.
---

# Blog Engine

## Use it for

- Blog planning
- Article briefs and outlines
- Writing or editing articles
- Refreshing decayed content
- Fact-check and citation-readiness review

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Choose a reader problem and search or distribution job before choosing a title.
- Use evidence for factual claims and distinguish original analysis from sourced facts.
- Do not pad articles to hit word counts.
- When optimizing for AI citation, improve evidence and extractability without degrading human usefulness.
- Choose article structure from the evidence, reader question, and argument. Do not force every article into a generic problem-solution sequence.

## Workflow

1. Define audience, query or distribution context, article job, existing coverage, and evidence sources.
2. Choose angle and scope that add something useful beyond a generic summary.
3. Build a brief with intent, key questions, proof, examples, outline, internal links, and conversion role.
4. Write or edit for clear reasoning, concrete examples, and scan-friendly structure.
5. Run fact-check, SEO, and GEO/AEO checks as relevant.
6. Plan refresh triggers and repurposing.

## Output contract

- Strategy, brief, outline, article, or audit as requested
- Evidence and citation notes
- Internal-link and distribution suggestions
- Refresh or reuse plan when relevant

## Quality gate

- No unsupported factual claims
- No redundant sections
- The article gives the reader a usable answer
- SEO and GEO choices do not distort the article

## Human-writing pass

Before final delivery, preserve the author's real voice and factual detail while removing generic filler, vague attribution, manufactured profundity, repeated rhetorical structures, and explanatory clauses that add tone instead of evidence. Do not regularize every paragraph into the same shape. If a sentence could move unchanged to an unrelated article, replace it with subject-specific evidence or cut it.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: brand-context
description: Use when the user needs a reusable marketing context for a brand, product, client, campaign, or offer. Captures audience, pains, alternatives, differentiation, proof, objections, voice, goals, and constraints.
---

# Brand Context Builder

## Use it for

- New client or brand onboarding
- Product marketing context
- Refreshing a stale brief
- Making later marketing outputs more consistent

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Capture only evidence the user supplied or that can be verified.
- Use customer language where available instead of corporate phrasing.
- Distinguish brand-level facts from campaign-specific choices.
- Do not force B2B persona fields onto consumer products or vice versa.

## Workflow

1. Summarize the product, business model, market, and current objective.
2. Define priority audience and buying context, including trigger, job, desired outcome, obstacle, and risk.
3. List real alternatives, including doing nothing, manual work, incumbent tools, and competitors.
4. Define differentiated value and the proof that supports each important claim.
5. Capture objections, anti-audiences, switching friction, voice rules, prohibited claims, and operational constraints.
6. Record the KPIs and decisions this context should support.

## Output contract

- A compact Brand Context document
- An evidence table for claims and proof
- A short list of missing inputs ranked by decision impact

## Quality gate

- Every major claim has a proof status
- Audience description includes behavior or job context, not demographics alone
- Voice guidance contains concrete do/don't examples
- The context is reusable across later skills

## Voice calibration from real samples

When the user supplies previous writing, treat it as stronger evidence than a generic style checklist. Infer recurring vocabulary, sentence rhythm, paragraph shape, opening and closing habits, level of directness, humor, uncertainty, punctuation, and patterns that are consistently absent. Preserve contradictions instead of smoothing them into a synthetic "perfect" voice. A voice rule should be traceable to the user's samples or an explicit preference.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: campaign-planning
description: Use when the user needs a campaign brief or integrated campaign plan with objective, audience, message, offer, channel roles, assets, budget logic, timeline, measurement, dependencies, and launch readiness.
---

# Campaign Planning

## Use it for

- Integrated campaigns
- Demand generation
- Seasonal or promotional campaigns
- Brand or product campaigns
- Turning a brief into executable channel work

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- One campaign must have one primary business objective.
- Keep channel roles distinct. Do not paste the same message everywhere.
- Use the offer and proof the brand can actually support.
- Plan measurement before launch.

## Workflow

1. Define objective, audience, trigger, offer, desired action, baseline, and constraint.
2. Choose the campaign idea and message hierarchy.
3. Assign each channel a role across discovery, consideration, conversion, or retention.
4. List required assets, landing experience, tracking, dependencies, approvals, and owners.
5. Create budget logic and pacing assumptions if budget exists.
6. Define launch checklist, in-flight review cadence, and post-campaign readback.

## Output contract

- Campaign brief
- Channel plan
- Asset and dependency matrix
- Timeline
- Measurement plan
- Launch-readiness checklist

## Quality gate

- No channel is included without a job
- Asset list matches platform and funnel needs
- Tracking and success criteria exist before launch
- Risks and dependencies are explicit

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: competitive-intelligence
description: Use when the user wants competitor research, category analysis, positioning comparison, offer comparison, content or channel gaps, battlecards, alternative pages, or a market narrative scan.
---

# Competitive Intelligence

## Use it for

- Competitor profiling
- Category and narrative mapping
- Offer, pricing, message, channel, SEO, or content comparison
- Building battlecards or alternative-page evidence

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Compare competitors on the same criteria.
- Use current public evidence when the request concerns live competitor behavior.
- Do not infer revenue, spend, traffic, or performance without a source.
- Treat absence of evidence as unknown, not weakness.

## Workflow

1. Define the decision the comparison must inform.
2. Select a small set of direct competitors, substitutes, and the status quo when relevant.
3. Collect comparable evidence for audience, promise, proof, offer, pricing, experience, channels, and visible acquisition patterns.
4. Separate parity, meaningful advantage, unproven claim, and open territory.
5. Translate findings into specific strategic choices and watch items.

## Output contract

- Comparison matrix
- Narrative and claim map
- Opportunity list ranked by evidence and strategic fit
- Risks and items requiring verification

## Quality gate

- Every important observation points to evidence
- No vanity feature-counting
- Recommendations explain why the user can credibly own the territory
- Current claims include dates or source recency when material

## Sales battlecard mode

When the request is for a competitive battlecard, keep it usable during a sales conversation. Include the competitor's current position, the few dimensions that matter to the buyer, where the user's offer is stronger, where the competitor is stronger, evidence-backed objection responses, and observable win/loss patterns when data exists. Never invent weaknesses, customer quotes, funding, pricing, or win rates. Unknowns stay unknown.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: content-strategy
description: Use when the user needs a content strategy, editorial direction, topic architecture, content pillars, format mix, distribution plan, repurposing plan, calendar logic, or a way to decide what content to create.
---

# Content Strategy

## Use it for

- Content strategy
- Topic and pillar planning
- Editorial calendars
- Repurposing and distribution
- Content portfolio diagnosis

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Start from audience questions, buying moments, search demand, distribution, and business goals.
- A content pillar must earn its place through audience relevance and business fit.
- Do not confuse posting frequency with strategy.
- Plan distribution while deciding the content, not after it is written.

## Workflow

1. Define audience, business objective, funnel role, existing assets, channels, and production constraints.
2. Map recurring audience problems, decisions, questions, proof needs, and cultural moments.
3. Group topics into a small number of durable content territories.
4. Assign formats and channels based on how the audience consumes and shares the information.
5. Create a mix of evergreen, timely, proof-led, educational, and conversion-support content.
6. Design repurposing paths and measurement per content role.

## Output contract

- Content thesis
- Content territories and topic examples
- Format/channel matrix
- Editorial cadence or calendar logic
- Distribution and repurposing plan
- Measurement by content role

## Quality gate

- Each territory has an audience reason and business reason
- No filler calendar slots
- Distribution is specific
- Metrics match the content's job

## Content matrix mode

When the user needs idea generation at scale, cross a small set of durable content territories with meaningfully different content jobs or formats. Each cell must produce a specific idea that could only belong to that territory, not a reusable theme. Use the matrix to expose gaps and variety, then rank the strongest ideas by audience relevance, evidence available, distribution fit, and business role. Do not fill a calendar merely to make every cell or date non-empty.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: copywriting
description: Use when the user wants to write, rewrite, sharpen, or review marketing copy for landing pages, websites, ads, emails, product pages, sales pages, CTAs, scripts, or campaign assets.
---

# Marketing Copywriting

## Use it for

- New copy
- Rewrites
- Message-to-copy translation
- Headline and CTA work
- Copy review before publishing

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Clarity beats cleverness when the two conflict.
- Use customer language, concrete detail, and proof.
- Write for the reader's awareness and next decision, not for the company's org chart.
- Do not fabricate proof, urgency, scarcity, testimonials, or numbers.
- Avoid generic AI phrasing, filler transitions, repetitive sentence shapes, and empty adjectives.
- Choose the information and narrative structure from the material, reader, channel, and decision. Do not default to a generic problem-solution frame.

## Workflow

1. Identify the copy job, audience, awareness stage, source of traffic, offer, proof, objection, and desired action.
2. Choose one primary message and supporting evidence.
3. Draft the information hierarchy before polishing sentences.
4. Write specific copy using the customer's language and the channel's constraints.
5. Remove filler, claims without proof, duplicated ideas, and vague CTAs.
6. Run the result through `marketing-quality` when it will be public or client-facing.

## Output contract

- Finished copy in the requested format
- Optional alternatives only when they help a real decision
- Notes on assumptions or missing proof when material

## Quality gate

- First screen or first lines make the value clear
- Every strong claim has support or is softened
- CTA describes a real next action
- Copy sounds written for this audience, not a generic template

## Human-writing pass

For public-facing prose, run a final editing pass that preserves the writer's point and voice while removing generic model patterns. Prefer the minimum effective edit. Check for portable filler that could be pasted into another company unchanged, vague authority claims, fake-insight setups, repeated sentence shapes, forced binary contrasts, stacked dramatic fragments, inflated significance, decorative formatting, unsupported specificity, and conclusions that only restate the piece. Keep useful rough edges when they belong to the writer.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: creative-testing
description: Use when the user needs ad angles, hooks, creative briefs, test matrices, fatigue diagnosis, message tests, visual concepts, UGC directions, or a structured way to learn which creative idea drives better response.
---

# Creative Testing

## Use it for

- Paid creative briefs
- Hook and angle generation
- Creative test design
- Fatigue or performance diagnosis
- Turning customer insights into ad concepts

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Test meaningful variables, not cosmetic changes labeled as experiments.
- Every concept must connect audience tension, promise, proof, and format.
- Do not invent creator testimonials or product demonstrations.
- Keep production feasibility and asset rights in scope.

## Workflow

1. Define audience, buying state, offer, platform, current winner/loser signals, and creative constraints.
2. Generate distinct hypotheses across problem, outcome, proof, mechanism, objection, comparison, or demonstration.
3. Choose formats that can express each hypothesis clearly.
4. Build a test matrix with one primary learning goal per cell.
5. Specify production brief, landing-page message match, metric, minimum read period or evidence threshold, and next action.

## Output contract

- Creative hypothesis bank
- Prioritized test matrix
- Production-ready briefs
- Readback and iteration plan

## Quality gate

- Concepts are meaningfully different
- Tests answer a decision
- Proof is real or explicitly needed
- Metric matches the creative job

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: cro-funnel
description: Use when the user wants to improve landing pages, websites, forms, signup, checkout, onboarding, funnels, conversion paths, message match, trust, friction, or test priorities.
---

# CRO and Funnel

## Use it for

- Landing-page audits
- Funnel diagnosis
- Signup or checkout improvement
- Form and CTA review
- CRO experiment backlog

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Diagnose before prescribing.
- Prioritize clarity, relevance, proof, friction, and risk over visual decoration.
- Use analytics, recordings, research, or other behavior evidence when available.
- Do not claim a conversion lift before a valid test.

## Workflow

1. Define page/funnel job, traffic source, audience, desired action, baseline, and known drop-off.
2. Review value clarity, message match, information hierarchy, proof, objections, CTA, friction, trust, and device context.
3. Locate the highest-leverage uncertainty or bottleneck.
4. Propose changes as hypotheses with rationale and evidence.
5. Prioritize tests by expected impact, confidence, effort, and dependency.
6. Define instrumentation and decision criteria.

## Output contract

- CRO diagnosis
- Quick fixes versus testable hypotheses
- Prioritized experiment backlog
- Copy or structure recommendations
- Measurement requirements

## Quality gate

- Recommendations are tied to a specific friction or uncertainty
- No decorative-only changes presented as CRO
- High-risk changes require testing
- Measurement is defined

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: customer-research
description: Use when the user wants to understand customers, buyers, users, pains, triggers, objections, Jobs-to-Be-Done, buying language, or patterns in interviews, reviews, calls, communities, surveys, and support data.
---

# Customer Research

## Use it for

- Synthesizing interviews, reviews, sales calls, support tickets, or survey responses
- Planning desk research when first-party research is missing
- Building JTBD, objection, trigger, and language insights
- Preparing inputs for positioning or copy

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Keep direct evidence separate from interpretation.
- Do not fabricate quotes or frequency counts.
- Frequency is not enough: capture intensity, business impact, and buying relevance.
- If desk research is used, identify source type and recency.

## Workflow

1. Choose mode: synthesize supplied evidence or research public sources.
2. Extract situation, trigger, desired progress, current workaround, pain, consequence, objection, proof requirement, and exact language.
3. Cluster recurring patterns without erasing meaningful minority segments.
4. Rank themes by decision relevance using frequency, intensity, and commercial impact.
5. Translate findings into messaging implications, product questions, and research gaps.

## Output contract

- Top themes with evidence
- JTBD and trigger statements
- Objection and proof matrix
- Customer-language bank
- Research gaps and recommended next interviews or sources

## Quality gate

- Quotes remain exact when quoted
- No invented percentages
- Interpretations are labeled
- Findings change a marketing decision, not just describe the audience

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: email-lifecycle
description: Use when the user needs lifecycle email strategy, welcome, nurture, activation, onboarding, abandoned flow, post-purchase, retention, win-back, newsletter, cold outbound, segmentation, frequency, or sequence copy.
---

# Email Lifecycle

## Use it for

- Lifecycle strategy
- Automated sequences
- Newsletter planning
- Retention and win-back
- Cold outbound when appropriate

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Map email to a customer state and next useful action.
- Respect consent, preference, frequency, and deliverability constraints.
- Do not use deceptive subject lines or fake reply/forward patterns.
- Separate marketing lifecycle and cold outbound requirements when they have different compliance obligations.

## Workflow

1. Define audience state, trigger, desired behavior, offer/value, segmentation, timing constraints, and consent context.
2. Design sequence logic before writing individual emails.
3. Assign one job to each message and specify the next state.
4. Write subject, preview, body, proof, CTA, and fallback path.
5. Set frequency, suppression, exit conditions, and experiment ideas.
6. Define delivery, engagement, conversion, and downstream business metrics.

## Output contract

- Lifecycle map
- Sequence table
- Email copy when requested
- Segmentation and suppression logic
- Measurement and test plan

## Quality gate

- Each message advances a state
- No sequence exists only to increase send volume
- Consent and exit logic are clear
- Success is measured beyond opens

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: geo-aeo
description: Use when the user wants visibility in AI-generated answers, citations, answer influence, entity consistency, AI-search content structure, AEO/GEO audits, citation readiness, or monitoring across AI answer surfaces.
---

# GEO and AEO

## Use it for

- AI answer visibility strategy
- Citation-readiness audit
- Entity and narrative consistency
- Content changes for answer engines
- AI visibility measurement design

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Do not equate being cited with shaping the answer. Treat source visibility and answer influence as separate outcomes.
- Use current evidence because AI answer products and crawler behavior change quickly.
- Do not guarantee citations or inclusion.
- Optimize for evidence quality, extractability, entity clarity, and broader source presence rather than keyword stuffing.

## Workflow

1. Define the target questions and the business outcome behind them.
2. Check current answer surfaces and cited source patterns when tools are available.
3. Audit whether the brand has clear entities, reusable claims, primary evidence, quotable facts, and accessible supporting pages.
4. Assess content extractability: direct answers, definitions, comparisons, evidence, authorship, dates, and structure.
5. Identify off-site sources and entity consistency gaps that affect trust.
6. Create a monitoring plan that separates mentions, citations, claim contribution, recommendation inclusion, and narrative accuracy.

## Output contract

- GEO/AEO diagnosis
- Question-to-evidence map
- Content and entity actions
- Off-site presence actions
- Measurement framework for citation and answer influence

## Quality gate

- No invented AI visibility data
- Recommendations are evidence-led
- Citation and influence metrics are distinct
- Human usefulness remains intact

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: growth-experiments
description: Use when the user wants growth ideas, acquisition or retention experiments, growth loops, referral mechanics, activation ideas, experiment prioritization, or a disciplined way to test growth hypotheses.
---

# Growth Experiments

## Use it for

- Growth backlog
- Acquisition or activation experiments
- Retention or referral loops
- Prioritizing uncertain growth bets
- Reviewing experiment results

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Growth ideas must connect to a bottleneck or opportunity in the customer behavior model.
- Prefer experiments that create learning, not activity.
- Do not label routine execution as an experiment unless a hypothesis and decision rule exist.
- Promotion from test to standard practice requires readback evidence.

## Workflow

1. Define the growth equation or funnel stage and current constraint.
2. Generate hypotheses tied to acquisition, activation, retention, revenue, referral, or a compounding loop.
3. Specify mechanism, target segment, expected behavior change, and risk.
4. Prioritize by impact, confidence, effort, speed of learning, and reversibility.
5. Design the smallest valid test and measurement plan.
6. After results, decide stop, iterate, expand, or operationalize.

## Output contract

- Growth diagnosis
- Prioritized experiment backlog
- Experiment cards with hypothesis, change, metric, duration/evidence threshold, and decision rule
- Readback template

## Quality gate

- Each experiment answers a decision
- Metrics include guardrails
- No false precision on expected lift
- Successful tests still require operational feasibility

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: launch-gtm
description: Use when the user needs a product launch, feature launch, go-to-market plan, launch tier, readiness checklist, message house, channel mobilization, sales enablement, launch timeline, or post-launch readback.
---

# Launch and GTM

## Use it for

- Product and feature launches
- GTM planning
- Launch tier and scope
- Readiness review
- Post-launch learning

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Scale launch effort to the importance and evidence behind the release.
- A launch is a coordinated market event, not simply a list of posts.
- Do not create proof that the product has not earned.
- Include product readiness, support, sales, measurement, and operational dependencies where relevant.

## Workflow

1. Define launch objective, audience, product truth, release scope, availability, market timing, and business risk.
2. Classify launch tier and required coordination.
3. Build positioning, message hierarchy, proof, offer, and launch story.
4. Plan channel roles, assets, sales/support enablement, PR/community activity, and internal readiness.
5. Create launch readiness gate with blockers, owners, and go/no-go criteria.
6. Define launch-day monitoring and post-launch readback.

## Output contract

- Launch brief
- Tier and scope rationale
- Message house
- Channel/asset timeline
- Readiness checklist
- Post-launch measurement plan

## Quality gate

- Claims match shipped product
- Dependencies are visible
- Launch scope matches business importance
- There is a post-launch learning loop

## Market-entry and motion selection

Before building the launch calendar, decide where and how the product should enter the market:

- Choose an initial segment using evidence about urgency, buying ability, access, competitive fit, and reference value.
- Compare GTM motions such as inbound, outbound, paid acquisition, community, partnerships, account-based work, and product-led adoption by economics and buyer behavior rather than popularity.
- Pick a primary motion and only the supporting motions that reinforce it. State why other plausible motions are deferred.
- Revisit the motion mix when ACV, sales cycle, product maturity, team capacity, or channel evidence changes.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: marketing-psychology
description: Use when the user wants behavioral science, decision psychology, mental models, persuasion analysis, friction diagnosis, choice architecture, trust, motivation, or habit principles applied to a marketing problem.
---

# Marketing Psychology

## Use it for

- Diagnosing why people hesitate or drop off
- Improving message, offer, pricing, CRO, onboarding, or retention decisions
- Designing ethical choice architecture
- Explaining a behavioral pattern in customer behavior

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Use behavioral concepts as hypotheses, not magic laws.
- Explain the mechanism and the expected behavior change.
- Avoid dark patterns, deception, coercion, false scarcity, or manipulative defaults.
- Prefer a small number of relevant principles over a long list of biases.

## Workflow

1. Define the target behavior and current behavior.
2. Identify the main motivational, cognitive, trust, or friction barrier.
3. Select one to three relevant behavioral principles.
4. Translate each principle into a specific intervention.
5. Define what evidence would support or reject the hypothesis.
6. Design a test or observation plan when practical.

## Output contract

- Behavior diagnosis
- Principle -> mechanism -> intervention table
- Ethical risk check
- Measurement or experiment suggestion

## Quality gate

- The principle explains a real mechanism
- The intervention is concrete
- No unsupported certainty
- No deceptive persuasion pattern

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: marketing-quality
description: Use to review marketing work before it is published, spent against, presented to a client, or treated as a decision. Checks evidence, clarity, audience fit, claims, consistency, channel fit, execution readiness, measurement, and generic AI-style writing.
---

# Marketing Quality Gate

## Use it for

- Final campaign review
- Client deliverable review
- Copy and content QA
- Pre-launch gate
- Checking whether a recommendation is supported

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- The reviewer must be willing to fail the work.
- Do not hide unknowns behind confident language.
- Judge the output against its intended job and channel, not a generic style rubric.
- Blocking issues take priority over polish.

## Workflow

1. Restate the intended audience, objective, channel, decision, and evidence base.
2. Check evidence and claims: sourced, supported, current where needed, and not overstated.
3. Check strategy: one primary message, audience relevance, offer fit, and action clarity.
4. Check execution: format, channel constraints, production readiness, consistency, and accessibility where relevant.
5. Check writing for generic filler, repeated structures, vague abstractions, unsupported superlatives, and robotic transitions.
6. Check measurement: success definition, tracking, readback, and decision rule.
7. Return PASS, PASS WITH FIXES, or FAIL with blocking issues first.

## Output contract

- Verdict
- Blocking issues
- High-impact fixes
- Optional polish
- Evidence or measurement gaps

## Quality gate

- A PASS has no unresolved blocking issue
- Every fix points to a specific part of the work
- Style feedback does not override strategy
- The review does not invent missing evidence

## Human-writing audit

For copy, posts, articles, scripts, decks, and narrative sections, check clusters of generic model patterns rather than treating one word or punctuation mark as proof of AI authorship. Flag portable filler, vague attribution, inflated importance, fake-insight framing, repeated contrast formulas, forced groups of three, synonym cycling, robotic cadence, stacked fragments, decorative headings, and unsupported specificity. Preserve a real writer's quirks when they are clear and intentional.

## Decision red-team

For a strategy, launch, media plan, or major recommendation, identify the load-bearing assumptions. Review the strongest plausible version of each assumption before challenging it. Rank only the few failure modes that could change the decision, then attach a cheap evidence check and a clear decision threshold. Do not inflate the risk list to make the review look thorough.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: marketing-strategy
description: Use when the user needs a marketing direction, annual or quarterly plan, channel priorities, growth thesis, budget logic, objectives, funnel choices, or a strategy that connects market evidence to execution.
---

# Marketing Strategy

## Use it for

- Annual or quarterly planning
- Client strategy
- Channel prioritization
- Budget and resource tradeoffs
- Turning diagnosis into a coherent set of marketing choices

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Strategy means choices and exclusions, not a list of tactics.
- Tie recommendations to the business model, audience, stage, budget, and evidence.
- Do not estimate outcomes with false precision.
- Separate strategic bets from operating hygiene.

## Workflow

1. Define business objective, economic constraint, time horizon, and current baseline.
2. Diagnose audience demand, offer strength, competitive context, funnel constraints, and measurement quality.
3. Choose priority segments, message, acquisition motions, and retention/growth motions.
4. Define the few strategic bets that can change the outcome.
5. Assign channels by role in the buying process, not popularity.
6. Set KPI tree, leading indicators, review cadence, and stop/continue rules.
7. Create a phased action plan with owners or resource needs where known.

## Output contract

- Strategy on a page
- Priority bets and explicit non-priorities
- Channel role matrix
- KPI tree and review cadence
- 30/60/90-day action plan or requested horizon

## Quality gate

- Every tactic serves a stated strategic choice
- Budget logic is explained
- KPIs map to business outcomes
- The plan can survive removal of fashionable channels

## Opportunity sizing and focus

When the strategic decision depends on market size or where to enter first:

1. Define the market boundary before calculating anything: problem, customer type, geography, use case, and buying unit.
2. Triangulate opportunity size with at least two independent methods when data permits, such as a category-level estimate and a customer-count × spend model.
3. Label sourced inputs, modeled inputs, assumptions, and confidence separately. Do not turn uncertain assumptions into precise forecasts.
4. Narrow the initial focus by comparing segment pain, ability and willingness to buy, access/distribution, competitive winnability, reference value, and strategic fit.
5. Prefer a focused entry segment when resources are limited, but never invent an arbitrary market-share target as proof that the segment is "winnable."

## Strategy stress test

For high-stakes plans, identify the 3 to 5 assumptions that would invalidate the strategy if false. For each, state the failure condition, evidence that could be collected cheaply, a decision threshold, and the smallest useful test. Give credit to assumptions that already have strong evidence instead of manufacturing doubt.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: mena-localization
description: Use when marketing work needs adaptation for Egypt, Saudi Arabia, UAE, Qatar, GCC, or broader Arabic-speaking markets, including Arabic copy, dialect, code-switching, cultural fit, currency, CTA length, RTL, and market-specific context.
---

# MENA Localization

## Use it for

- Arabic marketing copy
- Egyptian or Gulf localization
- Bilingual campaigns
- Ad, landing-page, email, or social adaptation
- Reviewing copy that sounds translated

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Localize the job of each line, not its words.
- Do not default every market to Modern Standard Arabic. Choose register by audience, channel, brand, and context.
- Keep brand names and technical/business terms in English when that is how the target audience naturally uses them.
- Flag legal, regulated, or culturally sensitive material that requires market-specific verification.
- Account for RTL and mixed Arabic/English layout constraints.

## Workflow

1. Identify target market, audience, channel, brand register, and desired action.
2. Extract the purpose of each source line: hook, proof, objection, instruction, or CTA.
3. Rewrite for local naturalness, cultural relevance, and placement length.
4. Adapt currency, examples, social proof, idioms, dates, and commercial context only when supported.
5. Flag items that should remain untranslated and any cultural or compliance risk.
6. Review the final copy aloud for natural rhythm and remove literal translation artifacts.

## Output contract

- Localized marketing copy
- Market/register note
- Flags for non-localized terms or risky references
- RTL or mixed-language production notes when relevant

## Quality gate

- Copy sounds native to the chosen market and brand
- Meaning and marketing job are preserved
- No unsupported local claim
- Layout constraints are considered

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: offer-pricing
description: Use when the user wants to design or improve an offer, package, price, guarantee, bonus, trial, promotion, value framing, monetization model, or pricing test.
---

# Offer and Pricing

## Use it for

- Offer construction
- Pricing and packaging
- Promotion design
- Trial or guarantee decisions
- Monetization experiments

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Treat price as an economic and positioning decision, not only a copy problem.
- Separate value hypothesis from willingness-to-pay evidence.
- Do not invent competitor pricing or customer willingness to pay.
- Avoid fake scarcity and misleading reference prices.

## Workflow

1. Define segment, job, alternative, current price context, cost constraints, and commercial goal.
2. Specify core outcome, included value, proof, risk reversal, and friction.
3. Design package structure and what changes across tiers.
4. Choose price metric and anchoring logic that fit how customers receive value.
5. Identify objections and what evidence or guarantee can address them.
6. Define a test plan for price, package, or offer changes.

## Output contract

- Offer architecture
- Pricing/package options with tradeoffs
- Objection and proof map
- Test plan

## Quality gate

- Each package has a clear audience and reason to exist
- Pricing logic matches value delivery
- No unsupported savings claim
- Test has a decision rule

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: paid-media
description: Use when the user needs paid-media strategy, account review, budget allocation, campaign structure, audience strategy, bidding logic, Google Ads, Meta Ads, LinkedIn Ads, TikTok Ads, creative planning, pacing, or spend diagnosis.
---

# Paid Media

## Use it for

- Paid-media planning
- Read-only performance review
- Budget and pacing decisions
- Campaign structure and targeting
- Cross-platform test planning

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Never assume live account access or current platform capabilities.
- For current policies, inventory, targeting, bidding, attribution, or product changes, verify current primary sources when tools are available.
- Separate platform-reported results from incrementality and business outcomes.
- Do not recommend spend changes from weak evidence without stating uncertainty and safeguards.

## Workflow

1. Define business goal, conversion event, economics, audience, market, budget, existing data, and tracking confidence.
2. Diagnose measurement quality before interpreting platform performance.
3. Assign platform and campaign roles based on demand state and available creative.
4. Design campaign structure, audience approach, bidding hypothesis, budget/pacing, and exclusions.
5. Define creative hypotheses and landing-page message match.
6. Set review cadence, thresholds, and readback fields before spending or changing spend.

## Output contract

- Paid-media plan or audit
- Budget/pacing logic
- Campaign and audience structure
- Creative and landing hypotheses
- Measurement and decision rules
- Risks and verification needs

## Quality gate

- No fake benchmarks
- Economics and tracking confidence are visible
- Recommendations have a reason and a decision rule
- No hidden assumption that platform attribution equals true incrementality

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: positioning-messaging
description: Use when the user needs positioning, value proposition, message hierarchy, message house, category framing, audience-specific messaging, proof architecture, or a sharper reason to choose the product.
---

# Positioning and Messaging

## Use it for

- New positioning
- Repositioning after product or market change
- Message hierarchy for campaigns or websites
- Turning research into language the market can understand

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Start from customer truth, real alternatives, and proof.
- A positioning statement is an internal decision aid, not necessarily final public copy.
- Do not claim uniqueness unless evidence supports it.
- Keep one primary message per audience and buying situation.

## Workflow

1. Select the priority segment and buying situation.
2. Name the real alternative and why it remains attractive.
3. Define the customer problem, consequence, trigger, and desired progress.
4. Identify differentiated value that matters to this segment and attach proof.
5. Write an internal positioning statement.
6. Convert it into message hierarchy: primary promise, supporting messages, proof, objection handling, and CTA direction.
7. Pressure-test for clarity, credibility, relevance, and competitive defensibility.

## Output contract

- Positioning statement
- Message hierarchy
- Proof and objection matrix
- Channel adaptation notes

## Quality gate

- The message is specific to a segment and situation
- Proof is attached to claims
- The language could be understood without product jargon
- The hierarchy can guide copy across channels

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: prepilot-router
description: Use for any broad or multi-part marketing request. Classifies the job, selects the smallest useful set of PrePilot marketing skills, sequences them, and keeps evidence, execution, QA, and measurement aligned.
---

# PrePilot Marketing Router

## Use it for

- Broad marketing requests with unclear method
- Multi-channel campaigns or launch work
- Requests that combine research, strategy, production, and measurement
- Cases where the user asks PrePilot what to do next

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Route by the user's job, not by marketing jargon in the prompt.
- Use one primary skill whenever possible. Add a second or third skill only when the job genuinely crosses stages.
- Do not create a long discovery questionnaire. Recover context from the conversation first, then ask only for missing inputs that materially change the answer.
- For execution work, route through marketing-quality before calling the output final when claims, spend, public publishing, or client delivery are involved.

## Workflow

1. Classify the request into one or more jobs: understand, decide, make, improve, launch, measure, or localize.
2. Identify the decision that must be made and the evidence required to make it.
3. Select the primary skill from the routing map below.
4. If work spans stages, sequence skills as evidence -> decision -> production -> quality -> measurement.
5. State major assumptions only when they affect the recommendation. Proceed with reasonable defaults for minor gaps.
6. Return the requested deliverable, not a description of the routing process.

## Output contract

- A focused answer using the selected workflow
- Visible assumptions only where material
- A short measurement or validation step when the output will be acted on

## Quality gate

- No more than three specialist skills for a normal request
- No duplicate analysis across skills
- No fake live data
- No vague next step such as "improve content" without specifying what to change

## Routing map

| User job | Primary skill | Common companion |
|---|---|---|
| Build reusable brand context | `brand-context` | `customer-research` |
| Understand customers | `customer-research` | `positioning-messaging` |
| Analyze competitors or category | `competitive-intelligence` | `positioning-messaging` |
| Define positioning or message hierarchy | `positioning-messaging` | `marketing-quality` |
| Decide the marketing direction | `marketing-strategy` | `analytics-measurement` |
| Build a campaign | `campaign-planning` | `creative-testing` |
| Apply behavioral science | `marketing-psychology` | the execution skill |
| Design offer or pricing | `offer-pricing` | `cro-funnel` |
| Write marketing copy | `copywriting` | `marketing-quality` |
| Build content plan | `content-strategy` | `social-content` or `blog-engine` |
| Create social work | `social-content` | `marketing-quality` |
| Plan or write blog content | `blog-engine` | `seo` or `geo-aeo` |
| Improve organic search | `seo` | `geo-aeo` |
| Improve AI answer visibility | `geo-aeo` | `seo` |
| Plan or review ads | `paid-media` | `creative-testing` |
| Build a creative test plan | `creative-testing` | `paid-media` |
| Improve conversion | `cro-funnel` | `copywriting` |
| Build lifecycle email | `email-lifecycle` | `copywriting` |
| Find growth bets | `growth-experiments` | `analytics-measurement` |
| Define measurement | `analytics-measurement` | any execution skill |
| Plan a launch | `launch-gtm` | `campaign-planning` |
| Adapt for Egypt/GCC/MENA | `mena-localization` | the source execution skill |
| Review work before use | `marketing-quality` | the producing skill |
| Size TAM/SAM/SOM or choose entry segment | `marketing-strategy` | `customer-research` |
| Choose GTM motion or beachhead | `launch-gtm` | `marketing-strategy` |
| Build a competitive battlecard | `competitive-intelligence` | `positioning-messaging` |
| Define a North Star Metric | `analytics-measurement` | `marketing-strategy` |
| Calibrate voice from writing samples | `brand-context` | `copywriting` or `social-content` |
| Humanize or de-slop marketing prose | `marketing-quality` | the producing skill |
| Review a LinkedIn profile | `social-content` | `positioning-messaging` |

For full-funnel tasks, default to: `brand-context` or existing context -> `marketing-strategy` -> channel skill -> `marketing-quality` -> `analytics-measurement`.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: seo
description: Use when the user needs SEO strategy, keyword research, content gaps, on-page SEO, technical SEO diagnosis, internal linking, site architecture, local SEO, programmatic SEO, schema planning, or organic-search prioritization.
---

# SEO

## Use it for

- SEO audits
- Keyword and intent research
- Content clusters and site architecture
- On-page or technical diagnosis
- Local and programmatic SEO planning

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Verify current search guidance and product behavior when the answer depends on recent changes.
- Separate technical blockers, relevance, quality, authority, and SERP competition.
- Do not promise rankings.
- Do not create thin pages simply to scale keyword coverage.

## Workflow

1. Define business goal, target market, site type, conversion path, and current organic baseline if available.
2. Map search demand by intent and business relevance.
3. Assess current pages, gaps, cannibalization, internal linking, and technical constraints.
4. Prioritize work by impact, confidence, effort, and dependency.
5. Specify page-level changes and technical requirements.
6. Define measurement using impressions, clicks, qualified traffic, conversions, and business outcomes.

## Output contract

- SEO diagnosis
- Priority roadmap
- Keyword/intent or page map when relevant
- Technical and on-page actions
- Measurement plan

## Quality gate

- Recommendations tie to search intent and business value
- No bulk-content recommendation without quality safeguards
- Technical issues include verification method
- Priority order is explicit

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1

View saved version →

---
name: social-content
description: Use when the user wants social media strategy or content for LinkedIn, X, Instagram, TikTok, Facebook, YouTube Shorts, or similar channels, including hooks, post concepts, scripts, calendars, and repurposing.
---

# Social Content

## Use it for

- Social strategy
- Post or short-video creation
- Founder or brand thought leadership
- Content calendars
- Repurposing long-form work into social assets

## Operating rules

- Use context already present in the conversation before asking for more input.
- Separate known facts, reasonable assumptions, and unknowns. Do not present assumptions as evidence.
- When the answer depends on current platform rules, market conditions, pricing, benchmarks, or competitors, verify them with current sources when tools are available.
- Prefer concrete decisions, examples, and next actions over generic marketing advice.
- Do not invent campaign performance, customer quotes, research findings, testimonials, rankings, or competitor claims.
- Keep the requested market, language, funnel stage, audience awareness, budget, and channel constraints visible throughout the work.
- Respect platform-native behavior without chasing every trend.
- Write for a specific audience tension or useful idea.
- Avoid fake personal stories, invented anecdotes, or manufactured controversy.
- For current platform limits or formats, verify current rules when tools are available.

## Workflow

1. Define channel, audience, objective, creator/brand voice, available proof, and production constraints.
2. Choose content angles based on audience tension, insight, proof, or useful demonstration.
3. Match format to idea: text, carousel, short video, image, thread, or live interaction.
4. Create the hook, body structure, proof/example, and action.
5. Plan reuse across channels without copy-pasting the same asset.
6. Define what to measure for the post's actual job.

## Output contract

- Channel-ready content or calendar
- Hook and structure
- Production notes when needed
- Repurposing options
- Measurement note

## Quality gate

- The hook earns attention without baiting
- The body delivers the promised value
- Voice is consistent
- CTA is proportionate to the content

## Voice and performance modes

If the user provides 3 or more representative writing samples, calibrate voice from recurring signals and absences before drafting. Preserve distinctive vocabulary, rhythm, bluntness, humor, uncertainty, and recurring structures rather than polishing every post into the same style.

If the user provides historical post performance, compare the draft with the user's own high- and low-performing patterns. Use the data as evidence, not as a universal formula. Report the dataset, metric definition, sample size, and age of the data before drawing a conclusion. Never substitute another creator's benchmark and present it as the user's expected performance.

For LinkedIn profile requests, treat the profile as a conversion surface: clarify the desired action, audience, offer, proof, headline, about section, featured proof/links, and consistency with the user's content voice.

## Handoff

If the task is part of a larger marketing request, return the completed deliverable plus the evidence or decisions the next PrePilot skill needs. Do not repeat upstream analysis unless it changes the result.

Referenced files: 1