Prompt Alchemist Navigate & Excel in the Agentic Economy

Prompt Alchemist · USP

Pattern Matrix

Ranked by difficulty. Clustered by function.

Every pattern holds two unique ranks out of 35. The top number is its scalability rank (1 = most fixed, 35 = most composable). The bottom-left number is its difficulty rank (1 = easiest to learn, 35 = hardest). Cards flow left-to-right from easiest to hardest, grouped by function within each row.

Switch to Periodic Table view

Card anatomy

S#

Ab

Pattern Name

D#
← Scalability ← Difficulty

D1 = easiest · D35 = hardest

S1 = most fixed · S35 = most composable

Difficulty levels

  1. Beginner Works out of the box, minimal setup
  2. Intermediate Requires some design & configuration
  3. Advanced Needs system design & integration work
  4. Expert Complex orchestration & deep expertise

Scalability levels

  1. Fixed Hard-coded, single-purpose use
  2. Configurable Parameterised, adapts to context
  3. Extensible Plugin-friendly, grows with the system
  4. Composable Combines freely with other patterns

Color — Function

  • Perception
  • Reasoning
  • Memory
  • Action
  • Coordination
  • Evaluation
  • Governance
Beginner Difficulty ranks 1–7 7 patterns
7

Pr

Prompt Engineering

6
2

ZsF

Zero / Few-shot

1
6

CoT

Chain-of-Thought

5
3

Ctw

Context Window

2
4

Sa

Solo Agent

3
5

Fv

Format Validation

7
1

Pg

Prompt Guardrails

4
Intermediate Difficulty ranks 8–15 8 patterns
13

Rag

RAG

13
8

Fn

Function Call

11
14

Tc

Tool Chaining

14
12

Sp

Scratchpad

12
11

Sh

Sequential Handoff

10
10

Sc

Self-Critique

9
9

Cp

Checkpoint Approval

8
15

Ob

Audit Trail / Observ.

15
Advanced Difficulty ranks 16–26 11 patterns
16

Msr

Multi-source Retrieval

18
25

As

Active Sensing

24
19

Pt

Parallel Tool Exec.

19
17

ToT

Tree-of-Thought / ReAct

17
26

Ref

Reflection + Replan

25
18

Vs

Vector Store

16
24

Ep

Episodic Memory

26
21

Sw

Supervisor / Worker

22
22

Xc

Cross-agent Critique

20
23

Ae

Automated Eval Loop

23
20

Ie

Interrupt & Escalation

21
Expert Difficulty ranks 27–35 9 patterns
35

Amb

Ambient Awareness

35
29

Wax

World-affecting Exec.

30
28

Al

Autonomous Action Loop

31
30

Mc

Meta-cognition

27
32

Pwm

Persistent World Model

32
27

Pp

Peer-to-Peer Network

29
33

Swm

Swarm Coordination

33
34

Sil

Self-improving Loop

34
31

Cal

Constitutional Alignment

28
D1
difficulty increases →
D35

How to read this view

Top-left = easiest to start

Cards in the first row carry the lowest difficulty ranks (D1–D7). These patterns need no infrastructure — just a prompt or a simple call.

Bottom-right = expert territory

The last row holds the highest difficulty ranks (D27–D35). These patterns demand deep architectural investment, operational discipline, and ongoing tuning.

Color clusters = same function

Within each row, cards are sorted by function group so matching colors sit together. Trace a color across rows to see how that function scales in complexity.

Top number = scalability rank

The number in the top-right corner ranks the pattern's scalability from 1 (most fixed, hard-coded) to 35 (most composable, adapts to any context). High top number = high leverage at scale.

Bottom-left number = difficulty rank

The number in the bottom-left ranks how hard the pattern is to learn and use, from 1 (minimal setup) to 35 (deep expertise required). This is also the card's position in the left-to-right flow.

Hover for full detail

Hover any card to see the full description, both rank numbers with their level labels, and the function group — all in context.