Use when Prompt Engineering Library is needed to produce a reusable prompt library for social-media or digital-marketing work; use `caption-writer` when its narrower outcome is requested.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Um comando direto ignora o prompt de revisão. Verifique a origem antes de executá-lo.
Instruções da origem · Visualização somente leitura
name
prompt-engineering-library
description
Use when Prompt Engineering Library is needed to produce a reusable prompt library for social-media or digital-marketing work; use `caption-writer` when its narrower outcome is requested.
Use this skill when the requested outcome is specifically a reusable prompt library and the supplied brief falls within prompt engineering library.
Do Not Use When
Use caption-writer when its narrower output is the real deliverable; do not use this skill as a generic substitute.
Do not use it to publish, send, spend, alter a live account, or make unsupported legal, platform, performance, or certification claims.
Required Inputs
Artefact
Source/provider
Required?
If absent
Content brief, channel, audience, message, format and call to action
Requester or approved brief
Yes
Stop and request the missing decision context.
Brand voice, offer facts, constraints and approvals
Client source pack or authorised owner
Conditional
State assumptions; do not invent names, prices, results or approvals.
Performance, platform or research evidence used for claims
Traceable export, URL, document or named source
Conditional
Draft the narrowest reviewable version and flag the missing evidence.
Capability and Permission Boundaries
Drafting is permitted within the supplied brief. Publishing, sending, spending, changing live accounts, or claiming certification requires separate explicit authority. Minimum capabilities are read access to supplied files and search across the authorised evidence set. Use only the files, tools, accounts and evidence made available for the engagement, expose every unassessed check, and obtain explicit authority before any mutation.
Degraded Mode
Fallback: if files, network access, platform data, language review or production tools are unavailable, return the narrowest useful qualified reusable prompt library; mark unavailable checks not assessed and never convert them into a pass.
Decision Rules
Choice
Action
Failure or risk avoided
Channel, format and audience commitment level are known
Choose the hook, structure and call to action native to that context.
Copy that could be pasted unchanged onto any channel or brand.
A required fact or approval is missing
Stop that claim or action; request it or use an explicit placeholder.
Fabricated facts, implied consent or unauthorised publication.
Evidence is partial but a useful draft is possible
Deliver a qualified draft with gaps and the next verification step.
Treating an unassessed requirement as passed.
Workflow
Confirm the exact reusable prompt library, consumer, market, channel and approval boundary; route to caption-writer if it is the closer match.
Inventory supplied facts, source provenance, constraints and missing inputs; stop if the objective, audience or authority is unknowable.
Select the domain method and record the material decision behind it before drafting.
Produce the smallest complete reusable prompt library; keep facts traceable and placeholders visibly unresolved.
Test the result against the decision table, domain quality criteria and anti-slop gate; recover by narrowing or qualifying unsupported portions.
Deliver the artefact with evidence, assumptions, unassessed checks and the next approval or verification step.
Outputs
Artefact
Consumer
Observable acceptance condition
Reusable prompt library
Requester, client reviewer or delivery team
The reusable prompt library addresses the named audience and objective, records assumptions, and passes the skill's domain checks without invented facts.
Decision and gap note
Approver or next workflow
Names the chosen route, evidence used, unresolved inputs and any action requiring authority.
Evidence Produced
Evidence
Format
Acceptance condition
Source/assumption register and completed release checklist
Inline table, checklist or linked source note
Every material claim, decision and unavailable check is traceable.
Quality Standards
Preserve the domain guidance and East African market context below; replace it only when the requester names another market.
Use British English unless the target language or market requires otherwise, and verify names, figures, quotations and platform rules before use.
Make the key choice visible, cover failure and edge cases, and keep the result ready for its named consumer.
Run the repository's anti-ai-slop ship gate; a blocking factual, cultural, safety or permission defect stops release.
Anti-Patterns
Writing before the objective and audience are known. Fix: stop and obtain the missing brief fields.
Reusing a neighbouring skill's template because the headings look similar. Fix: route by the requested reusable prompt library, not vocabulary overlap.
Adding a price, result, quotation, platform limit or cultural claim without a traceable source. Fix: verify it or qualify/remove it.
Treating missing access, evidence or native-language review as approval. Fix: mark the check not assessed and narrow the result.
Publishing, sending, spending or changing a live account from drafting authority alone. Fix: obtain explicit action-specific authority and retain the approval record.
References
caption-writer is the nearest routing comparison for this skill.
Industry — sector and sub-sector (e.g. financial services / SACCO)
Country/City — defaults to Uganda/Kampala if not specified
Primary goal — the main business objective this library will serve (e.g. lead generation, brand awareness, community growth)
Brand Context Block — paste the output from brand-voice-ai-training, or provide: brand voice descriptors, banned vocabulary, audience description, and primary WhatsApp number
Active platforms — list the platforms the client publishes on
Priority tasks — which recurring tasks the team performs most often (captions, blogs, reports, etc.)
The Master Prompt Formula
Source: Upadhyay, S. (2024) Generative AI for Marketing.
Every prompt in this library follows the Alpha-Beta-Gamma-Delta-Epsilon structure. Always prepend the client's Brand Context Block before the formula.
[Brand Context Block]
Consider the situation [Alpha — context and situation: what is happening, why it matters now].
Acting as [Beta — the creator persona AND a description of the target audience].
Perform [Gamma — the specific task to be completed].
In [Delta — the output format: caption, table, document, bullet list].
Using [Epsilon — boundaries, tone, style, length, platform rules, hard constraints].
Alpha — sets the scene. Include the product or service being promoted, the audience's current emotional or practical context, and any timely trigger (season, event, campaign phase).
Beta — defines two roles simultaneously: who is writing (the AI persona) and who they are writing for (the target reader). Both must be specific.
Gamma — states the task with precision. Vague tasks produce vague output. Include the copywriting framework (PAS, AIDA, etc.) when relevant.
Delta — specifies the exact format. "A caption" is insufficient; "a single caption with a line break after the hook, body text, and a CTA on a new line" is correct.
Epsilon — imposes hard constraints: language standard, word count, banned words, required elements (CTA, hashtags, WhatsApp link), and anything the output must never do.
The 10 Prompt Components
Source: Upadhyay (2024). Every strong prompt draws on some or all of these components. Identify which are required for each task before writing the prompt.
#
Component
What it controls
1
Persona
The creator role (e.g. social media copywriter) and the target audience (e.g. urban Ugandan women aged 25–40)
2
Voice
Tonality, emotional register, and formality level (e.g. warm and conversational; authoritative but approachable)
3
Style
The copywriting framework to apply: PAS, AIDA, BAB, FAB, SSS, PPPP, or AFOREST
4
Parameters
Length (word count, character limit), keywords to include, structural requirements
5
Channel
The exact platform (Facebook, LinkedIn, WhatsApp broadcast, email, blog) — each has different norms
6
Output type
The specific deliverable: caption, thread, email, table, bullet list, report narrative
7
Subject/Objective
The topic being covered and the specific goal of the piece (awareness, enquiry, purchase, retention)
8
Actions
The call-to-action required — what the reader should do next and how
9
Reference
A URL, competitor post, or LinkedIn profile the AI should learn from or match in quality
10
Conditionality
Hard constraints: never start with a question; always use British English; exclude competitor names; cite sources
Copywriting Style Frameworks
Apply via Component 3 (Style). Choose the framework that matches the audience's mindset and the content goal.
Framework
Structure
When to use
PAS
Problem → Agitate → Solve
Complaint-heavy audiences; content that solves a specific pain point
AIDA
Attention → Interest → Desire → Action
Standard awareness-to-conversion sequence; new product announcements
Persuasive speeches; thought leadership articles; keynote scripts
Prompt Library — Templates by Task
Replace all content in square brackets with client-specific detail before use. The Brand Context Block is abbreviated as [BCB] below; always paste the full block.
1. Caption Writing
[BCB]
Consider the situation: [describe what the business is promoting and why it matters
to the audience right now — include the campaign phase or seasonal trigger].
Acting as a social media copywriter for [Brand] writing for [audience description:
age, location, aspiration, current pain point].
Write a [platform] caption using the [PAS / AIDA / BAB — choose one] framework.
In a single caption with a line break after the hook, body text, and a CTA on a
new line.
Using: British English; maximum 150 words; no banned vocabulary; end with a
WhatsApp CTA [wa.me/256XXXXXXXXX]; include [number] hashtags drawn from
[hashtag guidance or list]; never begin with "Are you".
2. Blog Post Brief
[BCB]
Consider the situation: [describe the target reader's search intent or problem —
what are they typing into Google or asking on WhatsApp right now].
Acting as a content strategist briefing a writer for [Brand].
Create a complete blog post brief including: SEO title (under 60 characters),
meta description (under 155 characters), primary keyword, 5 H2 headings,
one practical takeaway section, and a closing CTA.
In a structured markdown document.
Using: British English; 1,200–1,800 word target; first H2 must include the primary
keyword; all examples must be drawn from Uganda or East Africa; no Western-market
assumptions; link suggestion for one internal and one external source.
3. Email Subject Line and Preview Text
[BCB]
Consider the situation: [state the email campaign goal, the audience segment,
and the stage in the customer journey — e.g. re-engagement, post-purchase, onboarding].
Acting as an email marketer writing for [audience description].
Generate 5 subject line options and 5 matching preview texts.
In a table format with columns: Subject Line | Preview Text | Technique Used.
Using: subject line under 50 characters; preview text under 90 characters;
no clickbait or false urgency; techniques may include curiosity, benefit,
urgency, social proof, or personalisation; include at least one option with
a localised East African reference; British English throughout.
4. Audience Persona
Consider the situation: [business name, industry, and Uganda/East Africa market —
include any known data about current customers or target segment].
Acting as a market researcher building a detailed audience persona for [Brand].
Create a complete audience persona for [client's target segment name or description].
In a structured document with the following fields: Name (fictional), Age, Location
(specific EA city or town), Occupation, Monthly income (UGX), Education level,
Social platforms used, WhatsApp behaviour (how they use it for commerce and comms),
content formats they engage with, their primary problem, what they want to achieve,
objections to buying, and one direct quote that captures their mindset.
Using: only realistic Uganda/East Africa demographics and income ranges; cite
Facebook penetration data where relevant; no Western market assumptions; reference
Chaffey (2024) for audience segmentation methodology if applicable.
5. Social Media Strategy — Platform Selection Section
[BCB]
Consider the situation: [client's current social media presence, primary business
goal, competitive context in Uganda/EA, and budget constraints if known].
Acting as a social media strategist applying the RACE framework
(Reach/Act/Convert/Engage — Chaffey, 2024) for [Brand].
Produce the platform selection and channel roles section of a social media strategy.
In structured markdown: one summary table (Platform | Primary Role | Content Type |
Posting Frequency) followed by a rationale paragraph per recommended platform.
Using: Uganda/EA platform penetration data; WhatsApp as the primary owned channel
for customer communications; Facebook as the primary reach channel; recommend no
more than 3 platforms for a starter client; cite the POEM model
(Paid/Owned/Earned) when classifying channels.
6. Community Management Response
[BCB]
Consider the situation: [paste or describe the comment, DM, or WhatsApp message
verbatim — include the sentiment, platform, and whether it is public or private].
Acting as a community manager for [Brand] responding to a
[complaint / question / compliment / crisis comment].
Write a response that [acknowledges the issue and offers resolution /
answers the question directly / expresses genuine gratitude].
In a [public comment reply / private DM / WhatsApp message] of under 60 words.
Using: warm, respectful East African professional tone; move the conversation
to WhatsApp if resolution requires further discussion
[wa.me/256XXXXXXXXX]; never be defensive or dismissive; never delete a
complaint without resolution; British English throughout.
7. Monthly Performance Report Narrative
Consider the situation: [paste the month's key metrics — reach, impressions,
engagement rate, follower growth, link clicks, WhatsApp enquiries generated,
and any notable campaign results. Include the previous month's figures for
comparison where available].
Acting as a social media analyst writing an executive summary for
[client's management team — state their level of digital literacy].
Produce a 3-paragraph performance narrative covering: (1) overall summary
and standout metric; (2) what drove the best-performing content and what
underperformed; (3) one recommended action for next month.
In plain, jargon-free English formatted as three labelled paragraphs.
Using: British English; cite specific numbers in every paragraph;
no hollow phrases such as "robust performance" or "going forward";
conclude with a forward-looking recommendation framed as a SMART objective;
reference the RACE framework (Chaffey, 2024) if applicable.
PAO Matrix — Pre-Prompt Checklist
Source: Joseph (c.2023–2024). Before writing any social media content prompt, confirm all three parameters simultaneously:
Platform — which platform is this content for? (Each platform has different norms, formats, and audience expectations.)
Audience — who specifically is this for, defined by values and mindset, not just demographics? (e.g. "first-generation entrepreneurs who value financial independence" not "25–40-year-olds")
Objective — what is the specific objective? Choose one: awareness / engagement / conversion / retention.
Omitting any one parameter forces the AI to make assumptions that produce generic output. All three must appear explicitly in every social media content prompt, in the Alpha and Beta elements of the master formula.
Prompt Anatomy Reference
Source: GPT Penguin (2024). Five-component anatomy that maps directly onto the Alpha-Beta-Gamma-Delta-Epsilon formula:
GPT Penguin Component
Description
Maps to Formula Element
Act Instruction
The role and expertise level assigned to the AI (e.g. "Act as a senior social media copywriter")
Beta (creator persona)
Context
The situation, background, and why it matters now
Alpha
Task
The specific action to perform
Gamma
Constraints
Hard limits: length, tone, banned words, format rules
Use this five-component map as a diagnostic: if a prompt is producing poor output, identify which component is missing or underdeveloped.
Copywriting Formula Prompt Activation
Source: Mizrahi (2024). Naming the formula in the prompt is more effective than describing the desired structure in prose. The activation pattern:
Write [content type] using the [FORMULA NAME] framework ([brief description of the sequence]).
Examples:
Write a Facebook caption using the PAS framework (Problem → Agitate → Solve).
Write a LinkedIn post using the AIDA framework (Attention → Interest → Desire → Action).
Write a promotional WhatsApp broadcast using the FOMO framework (Fear Of Missing Out — create urgency around a time-limited opportunity).
Write a community post using the SMILE framework (Subscriber, Meaningful, Inspiring, Likeable, Educational — positive engagement-first content).
Never describe the structure in prose ("make it compelling and build to a call to action") when you can name the formula directly. Named formulas produce structurally correct output; prose descriptions produce approximations.
Emotional Resonance Pattern
Source: GPT Penguin (2024). Human-sounding copy requires explicit emotional instruction. Every prompt for brand content must specify:
The emotion to evoke — trust, urgency, nostalgia, excitement, relief, pride, curiosity. Name it explicitly.
The brand value to connect it to — e.g. "connect the urgency to the brand's commitment to making financial services accessible."
The instruction to avoid manufactured urgency — add to the Epsilon layer: "Do not use artificial scarcity or false countdown language."
Activation pattern:
Using: evoke [EMOTION] by connecting to [BRAND VALUE]; do not use manufactured urgency or artificial scarcity.
Placeholder Variable Syntax
Source: Wright (2025). Every prompt template must use {{double-brace}} placeholders for all variable inputs:
Why this matters: Prompts without placeholders are single-use. Prompts with placeholders are agency assets — reusable across every client engagement. A prompt library built with consistent {{double-brace}} syntax can be searched, shared, and updated systematically. Apply to all templates in this library.
Hallucination Management Gate
Source: Evelyn (2025, p.51); Mizrahi (2024). For any output making factual or statistical claims, include this instruction in the prompt:
Use web search to find the latest news and resources, and cite your sources.
Every factual claim, statistic, named entity, date, or product claim in AI output must be flagged for human verification before client delivery. Do not assume AI-generated facts are correct. The hallucination gate is not optional — it is a production standard for any content that makes claims of fact.
### Separator Syntax
Source: Evelyn (2025). Any prompt containing both instructions and material to process (a brief, content to rewrite, example posts) must use triple hash marks (###) to separate sections. This prevents the model from confusing instructions with content to process.
Structure:
[Instructions here]
###
[Material to process — e.g. a brief, a draft to rewrite, example posts to learn from]
###
Apply whenever a prompt includes: pasted client briefs, existing copy for rewriting, competitor examples, or reference posts. Never mix instructions and source material in continuous prose.
Prompting Techniques
Source: Upadhyay (2024). Apply these methods when a single prompt does not produce sufficient output quality.
Prompt trail — break a complex deliverable into logical steps. Validate the AI's thinking at each step before proceeding to the next. Example: generate the strategy outline first, confirm it, then ask for the full narrative.
Multiple versions — request three variations of the same output (e.g. "Write three versions of this caption using three different frameworks"). Review all three, then ask the AI to merge the strongest elements into a final version.
Iterative refinement — use precise follow-up instructions to improve a draft. Example: "Add a local Uganda example to the second paragraph" or "Shorten the CTA to under 10 words."
Chain-of-thought — ask the AI to walk through its reasoning before writing the final version. Example: "Before writing the caption, explain which copywriting framework you are choosing and why it suits this audience." Review the reasoning; correct it if wrong before accepting the output.
Template approach — when a prompt produces excellent output, save the exact prompt (with client variables noted in brackets) as a named template in the client's project folder. Name templates clearly: caption-facebook-PAS-v1, email-subject-reengagement-v2. Reuse and iterate; do not start from scratch each time.
MVOSSTE Prompting Workflow (Randazzo, 2024)
Use this sequence when developing full marketing strategies with AI assistance:
Step
Prompt focus
Mission
"Draft 3 mission statement options for [client] that reflect [values]…"
Vision
"Write a 5-year vision statement assuming [growth scenario]…"
Objective
"Generate 5 SMART marketing objectives for [goal] in [timeframe]…"
Situation
"Conduct a SWOT analysis for [client] in [industry] in [market]…"
Strategy
"Suggest 3 strategic options for achieving [objective], with trade-offs…"
Tactics
"List 10 specific social media tactics for [strategy] with [budget] budget…"
Execution
"Create a 30-day action plan with weekly milestones for [tactic]…"
Professional practice: Footnote every AI-generated output with the exact prompt used. This enables clients to audit, reproduce, and modify outputs — and protects the consultant if outputs are later questioned.
Top-of-Funnel (Attract) — first-time visitors; no prior relationship with the brand.
[BCB]
Consider the situation: {{prospect description}} has just discovered {{brand name}} for the first time. They do not know the brand. Acting as a social media copywriter for {{brand name}} writing for {{audience values description}}. Write a {{platform}} post that introduces the brand's core value proposition without selling — educate and intrigue first. In a single post of under 120 words. Using: British English; no promotional language; no price mentions; end with a curiosity question or an invitation to learn more.
Middle-of-Funnel (Nurture) — leads who have shown interest (followed, engaged, clicked).
[BCB]
Consider the situation: {{prospect description}} has interacted with {{brand name}} content before. They know who the brand is. Acting as a social media copywriter for {{brand name}} writing for an audience that is considering but not yet committed. Write a {{platform}} post that provides proof of value — a case study, testimonial, or behind-the-scenes detail. In a narrative post of 100–150 words. Using: British English; specific and credible detail (no vague claims); include one social proof element; end with a soft CTA: "Find out how we helped [type of client]."
Bottom-of-Funnel (Decision) — decision-stage prospects ready to act.
[BCB]
Consider the situation: {{prospect description}} is actively deciding whether to work with {{brand name}}. They have seen the brand before. Acting as a social media copywriter for {{brand name}} writing for a warm, ready-to-act audience. Write a {{platform}} post that makes a direct, specific offer with a clear CTA. In a concise post of under 100 words. Using: British English; one specific offer or guarantee; one frictionless CTA ("Message us on WhatsApp: {{WhatsApp link}}"); no watered-down language; genuine urgency only where a real deadline exists.
Before/After Prompt Comparison — Social Caption
Source: Chavaux (2025). The most effective technique for making the quality gap between weak and strong prompts tangible.
BEFORE — Weak Prompt:
Write a caption for a Ugandan coffee brand.
Typical output: Generic, Western coffee-culture language. No local context. No CTA. No emotional pull.
AFTER — Strong Prompt:
[BCB for {{brand name}}, a specialty Ugandan coffee brand targeting urban Kampala professionals aged 28–42 who value locally-produced quality]
Consider the situation: It is Monday morning. Our audience is starting their working week and already thinking about productivity. Acting as a social media copywriter for {{brand name}} writing for ambitious Kampala professionals who take pride in supporting Ugandan-grown products. Write a Facebook caption using the AIDA framework (Attention → Interest → Desire → Action) that positions our coffee as the professional's daily ritual. In a single caption with a line break after the hook, body text, and CTA on a new line. Using: British English; maximum 120 words; evoke morning energy and local pride; end with a WhatsApp CTA {{WhatsApp link}}; include 3 relevant hashtags; never begin with "Are you".
Typical output: Locally grounded, emotionally resonant, framework-structured, with a clear CTA.
Forward-Reasoning Strategy Prompt
Source: Wright (2025). For strategic deliverables, use this pattern to generate differentiated thinking:
Imagine {{client's desired future state — e.g. "{{brand name}} is the most trusted financial services brand in Kampala"}}. Work backward to identify the key decisions and unexpected moves that contributed to this success. What did the brand do differently from competitors? What risks did it take? What did it stop doing?
This pattern bypasses generic strategic advice and produces counter-intuitive, specific insights by anchoring the AI's reasoning at the outcome rather than the starting point. Use for strategy documents, brand positioning, and 3–5 year planning sections.
Curiosity Gap Technique
Source: Wright (2025). The gap between what the reader knows and what they want to know creates click compulsion — the psychological pull that drives opens, clicks, and continued reading. Apply by naming it explicitly in the prompt:
Write [content type] that uses an intrigue gap to [desired response — e.g. "make the reader want to find out the answer," "compel them to open the full post," "create anticipation for the next step"].
The curiosity gap works by revealing enough to make the reader aware of what they do not know, then withholding the resolution until they take the desired action (click, open, reply, scroll). It is distinct from clickbait because it delivers the promised insight — it never misleads.
Quality Criteria
Output from this skill meets the standard when:
Every prompt template strictly follows the Alpha-Beta-Gamma-Delta-Epsilon structure and is prefixed by the Brand Context Block — no exceptions.
Each template contains all 10 prompt components relevant to that task; components that do not apply are explicitly omitted, not left blank.
Prompts produce output that a client team member can use without editing the prompt itself — all variables are clearly marked in square brackets with a plain-English instruction.
The library covers every recurring task the client identified in the Required Input and includes at least one template per active platform.
All character limits, word counts, and format rules in the Epsilon layer are specific and measurable — no vague instructions such as "keep it short."
Uganda/East Africa context is embedded in every relevant template: correct platform hierarchy, WhatsApp CTA format, UGX income ranges, and EA cultural references where applicable.
The copywriting framework specified in each caption or content template is the correct match for the task's audience mindset and conversion goal, with a one-line rationale provided.
The completed library is delivered as a single structured markdown document, organised by task category, ready to be saved in the client's project folder and shared with their team.
References
Chavaux, L. (2025) — Before/After prompt comparison methodology.
Erné, R. (2024) — 5-Step Perfect Prompt framework: Role, Context, Task, Format, Constraints.
