prompt-engineering

Name: prompt-engineering
Author: sabahattink

27stars9forks

How to install prompt-engineering?

sabahattink/antigravity-fullstack-hq/prompt-engineering

$npx -y skills add sabahattink/antigravity-fullstack-hq --skill prompt-engineering

Installs into the current project.

Summary

Comprehensive prompt engineering framework for designing, optimizing, and iterating LLM prompts. Use when creating prompts, optimizing existing prompts, or improving AI instructions.

Files · 1

View on GitHub

SKILL.md

1	# Prompt Engineering
2
3	## Workflow
4
5	```
6	User Request
7	\|
8	+-- "Create a prompt" --> EXPLORATION PHASE
9	+-- "Optimize this prompt" --> OPTIMIZATION PHASE
10	+-- "Fix this issue" --> ANALYSIS PHASE
11	```
12
13	## Phase 1: Exploration
14
15	Before creating any prompt, understand:
16
17	- What task will this prompt accomplish?
18	- Who will use it?
19	- What does success look like?
20	- What are the constraints?
21
22	## Phase 2: Analysis
23
24	### Task Classification
25
26	\| Dimension \| Options \|
27	\|-----------\|---------\|
28	\| Complexity \| Simple vs multi-step \|
29	\| Output \| Creative vs analytical vs structured \|
30	\| Stakes \| High vs experimental \|
31
32	### Strategy Selection
33
34	\| Task Type \| Approach \|
35	\|-----------\|----------\|
36	\| Simple \| Direct instructions \|
37	\| Complex \| Chain-of-thought \|
38	\| Creative \| Role setting \|
39	\| Structured \| Format specs + examples \|
40
41	## Phase 3: Implementation
42
43	### Version 1 - Minimal
44	- Core instructions only
45	- Test basic functionality
46
47	### Version 2 - Enhanced
48	- Add examples
49	- Clarify ambiguities
50	- Add constraints
51
52	### Version 3+ - Optimized
53	- Refine wording
54	- Remove redundancy
55
56	## Key Techniques
57
58	### Role Setting
59
60	```
61	As an experienced code reviewer, analyze...
62	```
63
64	### Chain-of-Thought
65
66	```
67	Think step-by-step:
68	1. First, identify...
69	2. Then, analyze...
70	3. Finally, conclude...
71	```
72
73	### Few-Shot Learning
74
75	```
76	Example 1:
77	Input: "Great product"
78	Output: { "sentiment": "positive" }
79
80	Now analyze: "It was okay"
81	```
82
83	### Explicit Constraints
84
85	```
86	- Limit to 3 paragraphs
87	- Focus on technical aspects only
88	- Do not include pricing
89	```
90
91	## Prompt Template
92
93	```markdown
94	## Context
95	[Background information]
96
97	## Role (Optional)
98	You are a [ROLE] with expertise in [DOMAIN].
99
100	## Task
101	[Clear instruction]
102
103	## Constraints
104	- Constraint 1
105	- Constraint 2
106
107	## Output Format
108	[Format specification]
109
110	## Examples (Optional)
111	[Input/Output examples]
112	```
113
114	## Common Mistakes
115
116	\| Mistake \| Fix \|
117	\|---------\|-----\|
118	\| Vague instructions \| Be specific \|
119	\| No examples \| Add 1-2 examples \|
120	\| Too many rules \| Simplify \|
121	\| No format spec \| Define output structure \|

People who liked this also viewed

6 picks

Type

prompt-engineering

27stars9forks

View on GitHub

How to install prompt-engineering?

sabahattink/antigravity-fullstack-hq/prompt-engineering

$npx -y skills add sabahattink/antigravity-fullstack-hq --skill prompt-engineering

Installs into the current project.

Summary

Comprehensive prompt engineering framework for designing, optimizing, and iterating LLM prompts. Use when creating prompts, optimizing existing prompts, or improving AI instructions.

Files · 1

View on GitHub

SKILL.md

1	# Prompt Engineering
2
3	## Workflow
4
5	```
6	User Request
7	\|
8	+-- "Create a prompt" --> EXPLORATION PHASE
9	+-- "Optimize this prompt" --> OPTIMIZATION PHASE
10	+-- "Fix this issue" --> ANALYSIS PHASE
11	```
12
13	## Phase 1: Exploration
14
15	Before creating any prompt, understand:
16
17	- What task will this prompt accomplish?
18	- Who will use it?
19	- What does success look like?
20	- What are the constraints?
21
22	## Phase 2: Analysis
23
24	### Task Classification
25
26	\| Dimension \| Options \|
27	\|-----------\|---------\|
28	\| Complexity \| Simple vs multi-step \|
29	\| Output \| Creative vs analytical vs structured \|
30	\| Stakes \| High vs experimental \|
31
32	### Strategy Selection
33
34	\| Task Type \| Approach \|
35	\|-----------\|----------\|
36	\| Simple \| Direct instructions \|
37	\| Complex \| Chain-of-thought \|
38	\| Creative \| Role setting \|
39	\| Structured \| Format specs + examples \|
40
41	## Phase 3: Implementation
42
43	### Version 1 - Minimal
44	- Core instructions only
45	- Test basic functionality
46
47	### Version 2 - Enhanced
48	- Add examples
49	- Clarify ambiguities
50	- Add constraints
51
52	### Version 3+ - Optimized
53	- Refine wording
54	- Remove redundancy
55
56	## Key Techniques
57
58	### Role Setting
59
60	```
61	As an experienced code reviewer, analyze...
62	```
63
64	### Chain-of-Thought
65
66	```
67	Think step-by-step:
68	1. First, identify...
69	2. Then, analyze...
70	3. Finally, conclude...
71	```
72
73	### Few-Shot Learning
74
75	```
76	Example 1:
77	Input: "Great product"
78	Output: { "sentiment": "positive" }
79
80	Now analyze: "It was okay"
81	```
82
83	### Explicit Constraints
84
85	```
86	- Limit to 3 paragraphs
87	- Focus on technical aspects only
88	- Do not include pricing
89	```
90
91	## Prompt Template
92
93	```markdown
94	## Context
95	[Background information]
96
97	## Role (Optional)
98	You are a [ROLE] with expertise in [DOMAIN].
99
100	## Task
101	[Clear instruction]
102
103	## Constraints
104	- Constraint 1
105	- Constraint 2
106
107	## Output Format
108	[Format specification]
109
110	## Examples (Optional)
111	[Input/Output examples]
112	```
113
114	## Common Mistakes
115
116	\| Mistake \| Fix \|
117	\|---------\|-----\|
118	\| Vague instructions \| Be specific \|
119	\| No examples \| Add 1-2 examples \|
120	\| Too many rules \| Simplify \|
121	\| No format spec \| Define output structure \|

People who liked this also viewed

6 picks

Type