Skip to content

Instantly share code, notes, and snippets.

@soltrinox

soltrinox/Unified-Prompt-Syntax-Guide.md

Created July 16, 2025 06:34
Show Gist options
  • Save soltrinox/f1be877041e8d527139cbf94bf1db31f to your computer and use it in GitHub Desktop.
Save soltrinox/f1be877041e8d527139cbf94bf1db31f to your computer and use it in GitHub Desktop.
Download ZIP
🛠️ Unified Prompt-Syntax Guide
Raw
Unified-Prompt-Syntax-Guide.md

🛠️ Unified Prompt-Syntax Guide

A hierarchical taxonomy and step-by-step architecture for high-precision ChatGPT sessions.

0. Design Philosophy

Teleology first, constraints second.
A prompt is a build script for reasoning. Declare the desired artefact; parameterize the dialectic (Λ); bind constraints; stage the workflow; embed self-checks.

1. Taxonomy of Intent → Hierarchy of Tactics

Tier Prompt Intent (𝚰) Typical Deliverables Primary Λ-Tactics
𝚰₁ Inform Facts, definitions, citations κ (audience precision), φ (tone neutrality)
𝚰₂ Analyze IRAC memos, comparative tables τ (format switch), δ (self-audit rubric)
𝚰₃ Create Stories, pitch decks, white papers γ (chunk by DAG), ε (contrast drafts)
𝚰₄ Transform Rewrites, translations, code refactors θ (Δ-prompts), κ (tool signals)
𝚰₅ Critique / Iterate Peer reviews, redlines, meta-feedback ζ (MWEs), ε (side-by-side deltas)

Hierarchy rule: higher tiers (𝚰₃-𝚰₅) usually nest lower ones—e.g., creation often embeds analysis and information retrieval.

2. Core Prompt-Craft Strengths (retain & amplify)

  1. α — Clarity of Objective
  2. β — Contextual Saturation
  3. γ — Structured Constraints & Style Guides
  4. δ — Domain-Specific Rigor
  5. ε — Iterative Mindset
  6. ζ — Self-Reflection & Meta-Queries

These strengths form the foundational layer; every advanced syntax element merely refines them.

3. Λ-Dialectic: Parameter Matrix

$$ \Lambda = (\kappa_{\text{audience}},; \sigma_{\text{verbosity}},; \tau_{\text{format}},; \phi_{\text{tone}}) $$

3.1 κ — Audience Selector

κ₁ Layperson • κ₂ Exec • κ₃ Board/Investor • κ₄ Counsel • κ₅ Regulator • κ₆ Domain Expert • κ₇ Engineer • κ₈ PM • κ₉ Academic • κ₁₀ Sales • κ₁₁ Press • κ₁₂ Internal Memo

3.2 σ — Verbosity Dial

σ₀ Flash ≤50w • σ₁ Brief ≈150w • σ₂ Compact ≈400w • σ₃ Standard ≈800w • σ₄ Expanded 1.5–2 k • σ₅ Comprehensive ∞

3.3 τ — Format Switch

τ₁ Bullets • τ₂ Numbered • τ₃ Outline • τ₄ IRAC • τ₅ Problem-Solution-Benefit • τ₆ Issue-Rule-Analysis-Conclusion • τ₇ FAQ • τ₈ Table • τ₉ Narrative • τ₁₀ Pseudocode • τ₁₁ Proof • τ₁₂ Slide Notes

3.4 φ — Tone Modulator

φ₁ Neutral • φ₂ Concise-Exec • φ₃ Board-Room • φ₄ Academic • φ₅ Technical • φ₆ Persuasive • φ₇ Conversational • φ₈ Thought-Leadership • φ₉ Journalistic • φ₁₀ Legalistic • φ₁₁ Storytelling • φ₁₂ Critical

4. Prompt-Syntax Architecture

GOAL:

Λ: κ = σ = τ = φ = θ_spec = <0–1 speculation tolerance> λ =

CONSTRAINTS:

  • style / vocabulary rules
  • banned words [...]
  • citation requirements
  • tool directives (python_user_visible, image_gen, ...)

STAGES:

  1. outline → freeze
  2. section drafts (G_i)
  3. integrated polish

[SELF-CHECK]

  1. Claims cite primary authority? □
  2. Banned lexemes absent? □
  3. Equations annotated inline? □

Ordering principle: Teleology → Λ → Constraints → Stages → Self-Check. Each layer refines the search space for the assistant’s reasoning engine.

5. Advanced Playbook (Greek-Indexed Techniques)

Symbol Technique Purpose
α Teleology Declaration Anchor the end-state.
β Λ-Parameterization Deterministic branching.
γ Chunk by Dependency Graph Prevent global rewrites.
δ Evaluation Hooks Enforce quality gates.
ε Contrast Prompts Surface trade-offs fast.
ζ Minimal Working Examples Style fidelity.
η Certainty Tolerance Control speculation.
θ Δ-Prompts Fine-grain revisions.
ι Time-Box Explorations Limit breadth, boost depth.
κ Tool Signals Route to proper execution channel.

6. Illustrative Scenario (Investor One-Pager)

GOAL: one-page investor brief introducing ENI6MA

Λ: κ = κ₃ # Board / Investor σ = σ₂ # ~400 words τ = τ₅ # Problem–Solution–Benefit φ = φ₂ # Concise-Exec θ_spec = 0.2 # Low speculation

CONSTRAINTS:

  • include TAM, SAM, SOM metrics
  • cite 1 external market report
  • banned words: "synergy", "milestone"

STAGES:

  1. outline → freeze
  2. draft each section
  3. polish & compress

[SELF-CHECK]

  1. Financial metrics present? □
  2. Report citation included? □
  3. ≤400 words total? □

7. Iterative Workflow Loop

  1. Draft 0 (outline)
  2. Freeze outline upon approval.
  3. Draft 1 (section fills) using Δ-prompts for edits.
  4. Self-Check executed automatically by assistant.
  5. Feedback → repeat Δ-cycle until acceptance.

8. Verbosity Control (λ)

If answers run long or short, set

$$ \lambda \in { \text{“ultra-brief”},; “brief”,; “normal”,; “expanded” } $$

The assistant scales detail logarithmically with λ.

9. Quick-Start Macro

Within 30 seconds you have a deterministic, reproducible path from prompt to polished artefact—no guesswork, no wasted iterations.

Below is a one-stop reference deck of every tuning knob mentioned so far, laid out as standalone markdown tables you can drop into any playbook or wiki.

1. Λ-Parameter Summary

Symbol Meaning Table Ref.
κ Audience §2
σ Verbosity / Length §3
τ Structural Format §4
φ Tone / Voice §5
θ Certainty Tolerance §6
λ Verbosity Override §7

2. κ — Audience Selector

Code Audience Archetype Prior Knowledge Example Artefacts
κ₁ Layperson / End-User Minimal How-to guides, FAQs
κ₂ Line-of-Business Executive KPI-driven Dashboards, ROI briefs
κ₃ Board Director / Investor Strategic & Fiscal Investor updates, one-pagers
κ₄ General Counsel / Attorney Legal nuance Bench memos, compliance notes
κ₅ Regulator / Auditor Statutory refs FedRAMP responses, audit packets
κ₆ Domain Expert (Cryptographer) Graduate-level tech Proof sketches, algorithm specs
κ₇ Software Engineer Code & APIs Run-books, API docs
κ₈ Product Manager User & market view PRDs, feature roadmaps
κ₉ Academic Reviewer Research methods Journal abstracts, lit reviews
κ₁₀ Marketing / Sales Lead Value props Pitch scripts, battle cards
κ₁₁ Press / Analyst Headline insights Press releases, analyst briefs
κ₁₂ Internal Memo Audience Org context Decision logs, policy drafts

3. σ — Verbosity Dial

Code Target Length Typical Uses
σ₀ ≤ 50 words (Flash) Taglines, tweets
σ₁ ~150 words (Brief) Elevator pitches, blurbs
σ₂ ~400 words (Compact) One-pagers, abstracts
σ₃ ~800 words (Standard) Blog posts, memos
σ₄ 1.5–2 k words (Expanded) White-paper sections, tutorials
σ₅ Unlimited (Comprehensive) Full reports, deep dives

4. τ — Structural Format Switch

Code Structure Best-Fit Scenarios
τ₁ Bullet List Quick scans, cheat sheets
τ₂ Numbered Steps Procedures, run-books
τ₃ Hierarchical Outline (I-A-1) Planning long docs
τ₄ IRAC Legal briefs
τ₅ Problem–Solution–Benefit Sales copy, proposals
τ₆ Issue–Rule–Analysis–Conclusion Rigorous legal/technical analysis
τ₇ FAQ Stakeholder question sets
τ₈ Comparison Table Feature/competitor matrices
τ₉ Narrative Paragraphs Case studies, storytelling
τ₁₀ Pseudocode / Algorithm Blocks Dev guidance
τ₁₁ Proof Outline Theorems, formal derivations
τ₁₂ Slide-Deck Notes Speaker cues for presentations

5. φ — Tone Modulator

Code Tone Label Stylistic Markers
φ₁ Neutral-Informative Objective, citation-focused
φ₂ Concise-Executive KPI-first, short sentences
φ₃ Board-Room Concise Formal brevity, risk framing
φ₄ Academic-Formal Passive voice, references, hedging
φ₅ Technical-Precise Jargon OK, specs & metrics
φ₆ Persuasive / Marketing Benefit-led, CTAs
φ₇ Conversational 1st/2nd person, approachable
φ₈ Thought-Leadership Visionary, trend-spotting
φ₉ Journalistic AP-style, inverted pyramid
φ₁₀ Legalistic Statutes, disclaimers
φ₁₁ Storytelling Anecdotes, vivid imagery
φ₁₂ Critical / Skeptical Counterpoints, risk emphasis

6. θ — Certainty Tolerance Scale

θ_spec Value Guidance to Model
0.0 Strict facts only
0.25 Low speculation; conservatively cite
0.50 Balanced facts & hypotheses
0.75 Exploratory; accept open questions
1.0 Blue-sky brainstorming welcome

7. λ — Verbosity Override Macro

λ Keyword Expansion Factor Practical Meaning
ultra-brief ×0.5 on σ Tagline-level condensation
brief ×0.75 on σ Executive summary
normal ×1 (default σ) Standard depth
expanded ×1.5 on σ Add detail & examples
exhaustive ×2+ on σ Comprehensive compendium

8. Advanced Playbook Symbols (α … κ)

Symbol Technique Purpose
α Teleology Declaration Anchor the end-state
β Λ-Parameterization Deterministic branching
γ Chunk by Dependency Graph Prevent global rewrites
δ Evaluation Hooks / Self-Check Enforce quality before delivery
ε Contrast Prompts Compare strategies side-by-side
ζ Minimal Working Examples Guarantee style fidelity
η Certainty Tolerance (θ_spec) Control speculative depth
θ Δ-Prompts Pin-pointed revisions
ι Time-Box Explorations Limit breadth, sharpen focus
κ Tool Signals (python_user_visible, etc.) Route to correct execution channel

9. Quick-Reference Cheat Grid

Λ Param Choose From Typical Command
κ κ₁ … κ₁₂ κ = κ₆
σ σ₀ … σ₅ σ = σ₂
τ τ₁ … τ₁₂ τ = τ₆
φ φ₁ … φ₁₂ φ = φ₅
θ 0.0 – 1.0 θ_spec = 0.25
λ keyword λ = expanded

Copy any row into your prompt block, adjust values, and the assistant will follow the specified branch with surgical precision.

🔚 End of Guide

📑 Prompt Examples for Five Use-Cases

(Copy → tweak → run)

1️⃣ Academic — Proof Outline for Journal Submission

GOAL: formal proof outline establishing NP-hardness of the Rosario–Wang exponential-search problem.

Λ:
  κ = κ₉        # Academic Reviewer
  σ = σ₄        # 1.5–2 k words
  τ = τ₁₁       # Proof Outline
  φ = φ₄        # Academic-Formal
  θ_spec = 0.25 # Low speculation
  λ = normal

CONSTRAINTS:
  - define every symbol on first use
  - cite ≥3 peer-reviewed PQC papers
  - banned words: “trivial”, “obvious”

STAGES:
  1) outline → freeze
  2) expand each lemma section
  3) integrate & polish

[SELF-CHECK]
1. All lemmas internally referenced?   □
2. Citations use IEEE format?         □
3. ≤2 000 words total?                □

2️⃣ Educational — Step-by-Step High-School Lesson

GOAL: numbered lesson plan teaching RSA fundamentals to Grade-12 students.

Λ:
  κ = κ₁        # Layperson / Student
  σ = σ₃        # ~800 words
  τ = τ₂        # Numbered Steps
  φ = φ₇        # Conversational
  θ_spec = 0.5  # Balanced facts & examples
  λ = brief

CONSTRAINTS:
  - include one hands-on activity
  - analogies must avoid advanced math
  - provide key-takeaway bullet list at end

3️⃣ Informative — Public-Facing FAQ

GOAL: FAQ explaining zero-knowledge proofs for a non-technical audience.

Λ:
  κ = κ₁        # Layperson
  σ = σ₂        # ~400 words
  τ = τ₇        # FAQ
  φ = φ₁        # Neutral-Informative
  θ_spec = 0.3  # Mostly factual

CONSTRAINTS:
  - 8–10 Q&A pairs
  - each answer ≤40 words
  - avoid the term “cryptographically robust”

4️⃣ Journalistic — Investigative Summary

GOAL: 500-word journalistic overview of recent NIST PQC standardization debates.

Λ:
  κ = κ₁₁       # Press / Analyst
  σ = σ₂        # ~400–500 words
  τ = τ₉        # Narrative Paragraphs
  φ = φ₉        # Journalistic
  θ_spec = 0.4  # Allow moderate context guesses
  λ = normal

CONSTRAINTS:
  - inverted-pyramid structure
  - quote at least one subject-matter expert
  - cite publication dates (MM-DD-YYYY) for all sources

5️⃣ Professional — Executive Strategy Memo

GOAL: 400-word memo outlining three strategic options for ENI6MA’s Q4 market expansion.

Λ:
  κ = κ₂        # Line-of-Business Executive
  σ = σ₂        # ~400 words
  τ = τ₅        # Problem–Solution–Benefit
  φ = φ₂        # Concise-Executive
  θ_spec = 0.2  # Low speculation
  λ = concise

CONSTRAINTS:
  - include one KPI per option
  - closing recommendation ≤50 words
  - banned words: “synergy”, “boil the ocean”

How to Use

  1. Paste the relevant block into ChatGPT.
  2. Edit the placeholders or constraints.
  3. Run — the model follows the Λ path and self-checks quality before delivery.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment