# Prompt Anatomy — full English prompt library

A practical AI playbook for teams and leaders: 5-part prompt framework, quick send check, copy-ready library, and a short quiz — less rework, more control.

Canonical lesson: https://promptanatomy.cloud/ · LT: https://promptanatomy.cloud/lt/

## Daily work

### lib_daySummary

Day summary from facts — fill the fields, get 5 bullets.

Input:
Your role (for context): [ROLE]
Today’s work: [TEXT]
Results: [TEXT]
Blockers: [TEXT]
Tomorrow’s plan: [TEXT]

Output:
5-bullet list
1–2 sentences each

### lib_taskList

From messy text — task list with priorities.

Input:
Text with ideas or requirements: [TEXT]

Output:
Task list
Priorities A / B / C
Deadlines if mentioned

### lib_meetingNotes

Turn rough notes into a structured meeting outcome.

Input:
Topic: [TOPIC]
Participants: [NAMES]
Notes: [TEXT]

Output:
Summary
Decisions
Actions (who / when)
Open questions

### mgr_daySummary

Leader day summary from facts — fill the fields.

Input:
Today’s focus and decisions: [TEXT]
Risks / blockers: [TEXT]
Team state (1–2 sentences): [TEXT]
Tomorrow’s priorities: [TEXT]

Output:
5-bullet leader summary (for execution; escalate to leadership if needed)
1–2 sentences each, minimal jargon

### mgr_taskList

From text — tasks with owner and priorities (for leaders).

Input:
Ideas, email, or meeting notes: [TEXT]

Output:
Task list with owners
Priorities A / B / C with rationale
Deadlines and dependencies if known

### mgr_meetingNotes

Meeting outcome for leadership or team — from short notes.

Input:
Topic and goal: [TOPIC]
Participants and roles: [LIST]
Short raw notes: [TEXT]

Output:
Summary for leader
Decisions and owners
Actions (who / by when)
Risks and open questions

## Communication

### lib_emailReply

Reply to an email or message you received.

Input:
Incoming email or message: [PASTE]

Output:
4–6 sentences
Professional, polite tone
Clear next step

### lib_simplifyText

Shorten text while keeping the core for the reader.

Input:
Text: [TEXT]
Audience: [AUDIENCE]

Output:
30–50% shorter version
Plain, clear language

### lib_hardUpdate

Hard but professional work message (deadline, error, change).

Input:
Channel: [EMAIL / SLACK / TEAMS]
Audience: [TEAM / CLIENT / PARTNER]
Facts: [WHAT HAPPENED / WHAT IS LATE / WHAT CHANGES]
Constraints: [WHAT YOU CANNOT SAY / CONFIDENTIALITY / DEADLINES]

Output:
2 variants: short (up to 120 words) and medium (up to 200 words)
Structure: context → what happens now → next step (who / when) → risks (up to 3)
Tone: calm, no blame; no speculation beyond facts

### lib_feedbackAfterMistake

Feedback after a mistake or incident (1:1, work context).

Input:
Relationship: [LEADER–MEMBER / PEERS]
Event (facts, no spin): [WHAT HAPPENED]
Impact: [TEAM / CLIENT / DEADLINES]
Goal for the conversation: [WHAT YOU WANT NEXT IN BEHAVIOR OR PROCESS]

Output:
Opening: 2–3 sentences (clear, not “school tone”)
2–3 observations tied to behavior and outcome
1 clear agreement or expectation
1 question to continue the dialogue
Avoid personal labels; talk about what can change

### lib_deescalateMessage

Rewrite a tense message to be calmer and clearer.

Input:
What caused tension: [BRIEF]
Current draft: [PASTE TEXT]
Recipient: [ROLE / AUDIENCE]

Output:
Rewritten message: facts → need → proposal → next step
Remove blaming phrases; keep a clear boundary if needed
If facts are missing — say what to ask instead of guessing

### mgr_emailReply

Reply for leadership or external comms.

Input:
Incoming email or message: [PASTE]

Output:
4–7 sentences
Tone: [formal / partner-like]
Clear next steps and ownership
Optional one line on confidentiality

### mgr_simplifyText

Shorten text for leadership, client, or market — one clear core.

Input:
Text: [TEXT]
Audience: [LEADERSHIP / CLIENT / MARKET / INTERNAL TEAM]
Goal (what the reader must grasp): [ONE STEP]

Output:
30–50% shorter version
1 sentence — core (what to do / what to know)
Plain language; no extra jargon or decoration
If political or legal risk remains — one clear note, no speculation

### mgr_hardUpdate

Hard message to leadership or outside — reputation and risk visible.

Input:
Channel: [EMAIL / SLACK / TEAMS]
Audience: [TEAM / CLIENT / PARTNER / LEADERSHIP]
Facts: [WHAT HAPPENED / WHAT IS LATE / WHAT CHANGES]
Risks: [LEGAL / COMMS / CONTRACT / REPUTATION — if any]
Constraints: [WHAT YOU CANNOT SAY / CONFIDENTIALITY / DEADLINES]

Output:
Two variants: short and medium
Structure: context → decision now → owner and deadline → risks (up to 3)
Tone: calm, no blame; no speculation beyond facts

### mgr_feedbackAfterMistake

Feedback to a leader: results and behavior (1:1).

Input:
Relationship: [LEADER–MEMBER / PEERS]
Event (facts): [WHAT HAPPENED]
Impact: [TEAM / CLIENT / DEADLINES / RISKS]
Goal: [WHAT YOU WANT NEXT IN BEHAVIOR / PROCESS]

Output:
Opening: 2–3 sentences (clear, not “grading” tone)
2–3 observations tied to behavior and impact
1 clear expectation and 1 visible next step
1 question to continue the dialogue
If HR / process step is relevant — one neutral note (no diagnoses)

### mgr_deescalateMessage

De-escalation for a leader message — assess risk and continuity.

Input:
What caused tension: [BRIEF]
Current draft: [PASTE TEXT]
Recipient: [ROLE / AUDIENCE]
Risk: [WHAT COULD BOOMERANG — if any]

Output:
Rewritten message: facts → need → proposal → clear next step
Remove blaming phrases; keep a clear boundary if needed
If facts are missing — say what to ask instead of guessing

## Quality check and constraints

### lib_qualityCheck

Review your prompt before high-stakes decisions or bulk sends.

Input:
[Paste your prompt here]

Criteria (5 principles): Clarity; Experimentation; Simple to complex; Context; Word choice.

Output:
1) Table: 1–5 score per principle
2) Briefly why that score
3) 3 weakest spots
4) Improved prompt version
Rules: be specific; if facts are missing — say so; if too complex — offer V1; do not change without upside.

### lib_constraints

Define an AI prompt with constraints so the output stays controllable.

Input:
Task (what the AI must do): [TASK]
Length: [E.g. max 200 words / 1 page]
Avoid: [E.g. jargon, hype, subjective ratings]
Format: [E.g. bullets, table, paragraphs]
Language: [LANGUAGE]

Output:
One clear prompt ready to paste into the AI
"Avoid": up to 5 short bullets

### mgr_qualityCheck

Review a prompt before bulk send or major leadership decisions.

Input:
[Paste your prompt here]

Criteria (5 principles): Clarity; Experimentation; Simple to complex; Context; Word choice.

Output:
1) Table: 1–5 score per principle
2) Briefly why that score
3) 3 weakest spots
4) Improved prompt version
Rules: be specific; if facts are missing — say so; if too complex — offer V1; do not change without upside.

### mgr_constraints

AI task with compliance and approval chain (for leaders).

Input:
Task or outcome: [WHAT MUST BE DONE]
Confidentiality: [INTERNAL / CLIENT / PUBLIC — what must not leak]
Legal / contract boundary: [WHAT IS FORBIDDEN OR NEEDS SIGN-OFF]
Length and format: [E.g. 1 page, bullets, table]
Language: [LANGUAGE]
Approval chain: [WHO SIGNS OFF BEFORE SEND — if relevant]

Output:
One clear AI task with context and format
"Avoid": up to 5 bullets
If facts are missing — what to ask before acting, not guessing

## Prompt patterns (general)

### lib_zeroShot

Write a one-line product tagline for [PRODUCT].

### lib_fewShot

Examples:
"[EXAMPLE 1]"
"[EXAMPLE 2]"
Write 3 more taglines in the same style for [PRODUCT].

### lib_thoughtChain

Identify 3 main problems related to [SITUATION / PRODUCT / PROCESS].
Suggest fixes.
Outline a short action plan.

### lib_promptSeq

Write 5 key talking points on [TOPIC].
Outline the structure.
Write a one-paragraph intro.

### lib_instruct

Write a [WORD COUNT]-word piece on [TOPIC].
Audience: [AUDIENCE].
Tone: [TONE].
Language: [LANGUAGE].

### lib_decisionSummary

META: You are a business analyst.
INPUT:
Situation: [DESCRIPTION]
Options: [OPTIONS]
Data: [NUMBERS / FACTS]
OUTPUT:
Problem
2–3 options
Recommendation
Risks

### lib_processDoc

META: You are a process specialist.
INPUT:
Topic: [PROCESS]
Current steps: [STEPS]
OUTPUT:
6–10 clear steps
Short sentences, verbs

### lib_swot

META: You are a business consultant.
INPUT:
Project / company: [DESCRIPTION]
Data: [FACTS]
OUTPUT:
SWOT table
3–5 concrete bullets per quadrant

### mgr_zeroShot

Write a one-line business tagline or positioning line for initiative / product [NAME]. Audience: leadership or external partners.

### mgr_fewShot

Examples (style for leadership or client):
"[EXAMPLE 1]"
"[EXAMPLE 2]"
Write 3 more lines in the same style on [TOPIC].

### mgr_decisionSummary

META: You are a leadership decision analyst.
INPUT:
Question or dilemma: [DESCRIPTION]
Options and arguments: [OPTIONS]
Data / facts: [NUMBERS / FACTS]
OUTPUT:
Core issue
2–3 scenarios with risk
Recommendation for leadership
What to watch after the decision

### mgr_processDoc

META: You are an org process and compliance expert.
INPUT:
Process or policy: [PROCESS]
Current practice: [STEPS / EXCEPTIONS]
OUTPUT:
6–10 clear steps (incl. owners and control points)
Short sentences, verbs
One note on audit or documentation if relevant

### mgr_swot

META: You are a strategy consultant for leadership.
INPUT:
Business unit / initiative: [DESCRIPTION]
Data and assumptions: [FACTS]
OUTPUT:
SWOT table
3–5 concrete bullets per quadrant
One note on competitive or regulatory context if relevant

## Other

### mgr_instruct

Write a [WORD COUNT]-word piece on [TOPIC].
Audience: [AUDIENCE].
Tone: [TONE].
Language: [LANGUAGE].

### mgr_promptSeq

Write 5 key talking points on [TOPIC].
Outline the structure.
Write a one-paragraph intro.

### mgr_thoughtChain

Identify 3 main problems related to [SITUATION / PRODUCT / PROCESS].
Suggest fixes.
Outline a short action plan.