✳ HEARTH & CODE / FIELDCRAFT

05 / PROMPT CRAFT / METHOD CATALOG

Prompt engineering catalog

Treat prompts as reusable methods with inputs, constraints, failure modes and return shapes — not as incantations.

The working catalog contains 128 prompt methods, with a broader 512-method horizon. The public layer shares the design of the catalog and synthesized examples, rather than exporting private prompt text or operational context.

Organized by transformation

Methods are grouped by the control mechanism they use, not by topic. Core techniques define the base operation — extract, compare, synthesize, evaluate, transform. Contrastive, Staged, Receipt, Boundary, Schema, Falsifier, Adaptive, and Position variants add different control structures to the same base technique.

This lets you choose a prompt by the transformation it performs rather than hunting for a topic label. Each category in the atlas covers one facet of prompt engineering — from instruction contracts to creative divergence to governance boundaries.

To use this catalog: Click any tile to see an 8-section deep dive: how the technique works, when and how to use it, limitations, related methods, and a worked example. Filter by control mechanism or search by name. The 128 techniques here are drawn from a 512-record research atlas.

How to read a technique card

Each technique card shows: name (what it's called in the atlas), control type (core / contrastive / staged / receipt / boundary / schema / falsifier / adaptive / position), and number (sequential within the 128-entry subset).

Tap or click any card to open the full 8-section pop-up: core mechanism, practical use, when/how to use it, limitations, related techniques, a synthesized example, and precise orientation with atlas metadata (PTA ID, category, document, evidence class).

Filter buttons at the top let you show all techniques or restrict to one control mechanism. The search box filters by name and description. Results animate on transition.

Sourcing and limits

These 128 records are drawn from the 512-record Prompt Technique Atlas — a structured corpus across 64 categories and 16 documents. Each record is a candidate (not independently verified for production) with named evidence class and source origin. The public catalog uses synthesized pedagogical examples, not operational prompt text.

Claim boundary: Catalog presence is not admission, endorsement, or deployment authorization. Each technique was developed for the task class described by its trigger; cross-domain transfer requires independent evaluation. Evidence classes range from survey-secondary to observed-hub-grounded — see each card's precise-orientation section for the exact PTA identifier.

128 RETAINED METHODS

Prompt Engineering Catalog

Each technique is organized by what it does to information, not what topic it addresses. Click any card to open its full details: how it works, when to use it, and what it cannot do.

Core#1

Dynamic Context Assembly

Contrastive#2

Contrastive-Fork Instruction Hierarchy and Precedence

Core#3

Emotion Prompting

Staged#4

Staged-Composition Task Framing and Intent Normalization

Core#5

Outcome-First Contract

Receipt#6

Receipt-Bound Achieved-State and Outcome Specification

Core#7

Bounded Receipt with Non-Effects

Contrastive#8

Contrastive-Fork Constraint and Non-Action Contracts

Core#9

Provenance-Carrying Chunk

Contrastive#10

Contrastive-Fork Source Grounding and Authority

Core#11

Evidence-per-Claim

Staged#12

Staged-Composition Claim-Evidence Binding

Core#13

Ambiguity Ledger

Staged#14

Staged-Composition Uncertainty Calibration and Missing Evidence

Core#15

Counterexample Set

Receipt#16

Receipt-Bound Counterevidence and Falsification

Core#17

Progressive Disclosure

Contrastive#18

Contrastive-Fork Context Selection and Progressive Disclosure

Core#19

KV-Cache Selective Retention

Staged#20

Staged-Composition Context Compression and Summarization

Core#21

Episodic Experience Memory

Boundary#22

Boundary-Bound Memory and Retrieval Design

Core#23

Staleness Gate

Staged#24

Staged-Composition Provenance, Currentness, and Staleness

Schema#25

Schema-Guided Structured Input Contracts

Staged#26

Staged-Composition Structured Input Contracts

Position#27

Position-Aware Context Packing

Staged#28

Staged-Composition Output Schemas and Typed Responses

Core#29

Structured Function Invocation

Receipt#30

Receipt-Bound Validation, Rejection, and Repair

Core#31

Positional Context Extension

Staged#32

Staged-Composition Transformation and Projection Contracts

Core#33

Goal-Prompt Lifecycle State Machine

Contrastive#34

Contrastive-Fork Agent Loops and State Machines

Core#35

Capability-Bounded Delegation

Staged#36

Staged-Composition Tool and Capability Boundaries

Core#37

Chain-of-Thought Prompting (Agentic Pattern)

Contrastive#38

Contrastive-Fork Budgets, Stops, and Attempt Accounting

Core#39

Context Snapshot and Restoration

Staged#40

Staged-Composition Recovery, Idempotency, and Continuity

Core#41

Wave Decomposition (12-wave / 15-wave / 8-pass naming)

Receipt#42

Receipt-Bound Task Decomposition and Wave Design

Core#43

Disjoint-Path Parallel Writers

Staged#44

Staged-Composition Dependency Graphs and Scheduling

Core#45

Checkpointed Autonomy

Receipt#46

Receipt-Bound Typed Handoffs and Returnability

Core#47

Multi-Tool Coordination

Contrastive#48

Contrastive-Fork Multi-Agent Coordination and Arbitration

Core#49

Evaluator-Optimizer Loop

Staged#50

Staged-Composition Rubrics and Multi-Dimensional Assessment

Core#51

Acceptance-Test Prompt

Receipt#52

Receipt-Bound Fixtures, Baselines, and Reproducibility

Core#53

Metamorphic Prompt Test

Contrastive#54

Contrastive-Fork Adversarial and Metamorphic Evaluation

Core#55

Cost-Latency-Quality Frontier

Staged#56

Staged-Composition Observability, Cost, and Latency

Core#57

Self-Refine Prompting

Receipt#58

Receipt-Bound Reflection and Self-Critique Prompts

Core#59

Human Decision Gate

Receipt#60

Receipt-Bound Preference Elicitation and Personalization

Core#61

Achieved-State-First Questioning

Core#62

Multi-Lens Option Framing

Core#63

Sparse or Linear Attention

Contrastive#64

Contrastive-Fork Immutable Return Intake and Disposition

Core#65

Retrieval-Augmented Generation

Receipt#66

Receipt-Bound Divergent Ideation

Core#67

Modular Retrieval Pipeline

Receipt#68

Receipt-Bound Recombination and Concept Blending

Core#69

Worked-Miniature Bridge

Receipt#70

Receipt-Bound Analogy and Cross-Domain Transfer

Core#71

Constraint Sandwich

Receipt#72

Receipt-Bound Constraint-Driven Creativity

Core#73

Retrieval-Augmented Generation (RAG)

Receipt#74

Receipt-Bound Novelty Search and Alternative Generation

Core#75

Layer-of-Thoughts Prompting

Receipt#76

Receipt-Bound Uniqueness and Local Specificity

Core#77

Demonstration-Conditioned Context

Receipt#78

Receipt-Bound Strange Options and Edge Cases

Core#79

Independent Parallel Sampling

Staged#80

Staged-Composition Anti-Convergence and Diversity Preservation

Core#81

Graph-Guided Retrieval

Receipt#82

Receipt-Bound Literature Discovery and Search Strategy

Core#83

Merge-Node Validation

Contrastive#84

Contrastive-Fork Research Synthesis and Claim Maps

Core#85

Agent Communication Protocol

Receipt#86

Receipt-Bound Methodology and Protocol Design

Core#87

Chain-of-Verification Prompting

Receipt#88

Receipt-Bound Citation and Source Audit

Core#89

Chain of Draft Prompting

Boundary#90

Boundary-Bound Decomposition and Intermediate Representations

Core#91

Contrastive Denoising with Noisy Chain-of-Thought

Receipt#92

Receipt-Bound Contrastive and Dialectical Reasoning

Core#93

Power-of-Two Batching

Receipt#94

Receipt-Bound Hypothesis Generation and Testing

Core#95

Failure Taxonomy Loop

Contrastive#96

Contrastive-Fork Causal and Systems Reasoning

Core#97

Workbook Instrument (~50-Question Reflection)

Receipt#98

Receipt-Bound Socratic Questioning

Core#99

Optimization by Prompting

Receipt#100

Receipt-Bound Explanation and Teaching Prompts

Core#101

Adaptive Agentic Retrieval

Staged#102

Staged-Composition Adaptive Scaffolding and Feedback

Core#103

Logical Chain-of-Thought Prompting

Receipt#104

Receipt-Bound Collaborative and Voice-Preserving Authoring

Core#105

Chain-of-Code Prompting

Contrastive#106

Contrastive-Fork Code and Software Engineering Prompts

Core#107

Structured Data Verbalization

Receipt#108

Receipt-Bound Data and Analytical Prompts

Core#109

Multimodal Context Integration

Receipt#110

Receipt-Bound Visual and Multimodal Prompts

Core#111

Structured Chain-of-Thought Prompting

Receipt#112

Receipt-Bound Scientific and Mathematical Prompts

Core#113

Context Manifest

Staged#114

Staged-Composition Privacy and Disclosure Boundaries

Core#115

Adversarial Review Gate

Staged#116

Staged-Composition Permissions and Consequential Effects

Core#117

Tree-of-Thoughts Prompting

Receipt#118

Receipt-Bound Policy and Constitutional Prompting

Core#119

Blind Comparator

Contrastive#120

Contrastive-Fork Audit, Receipts, and Governance Evidence

Core#121

Self-Harmonized Chain-of-Thought Prompting

Receipt#122

Receipt-Bound Meta-Prompting and Prompt Reflection

Core#123

Automated Prompt Search

Contrastive#124

Contrastive-Fork Prompt Generation and Optimization

Core#125

Buffer of Thoughts Prompting

Receipt#126

Receipt-Bound Workflow and Skill Compilation

Core#127

Iterative Self-Refinement

Receipt#128

Receipt-Bound Adaptive Automation and Self-Revision

KEEP THE CLAIM BOUNDED

The counts describe a working catalog and its broader horizon, not a claim that all methods are public, complete, equally mature or suitable for every model or context. Method cards on this page describe structural patterns; they do not reproduce operational prompt text. Registry entries are synthesized pedagogical examples unless labeled otherwise.