create-skill

Name: create-skill
Author: jinzhezenggroup

113stars23forks

How to install create-skill?

jinzhezenggroup/computational-chemistry-agent-skills/create-skill

$npx -y skills add jinzhezenggroup/computational-chemistry-agent-skills --skill create-skill

Installs into the current project.

Summary

Create new Agent Skills following the agentskills.io specification. Use when the user wants to create, scaffold, or design a new skill for AI agents. Handles SKILL.md generation, directory structure setup, and validation.

Files · 1

View on GitHub

SKILL.md

1	# Create Skill
2
3	This skill helps you create new Agent Skills that follow the [agentskills.io specification](https://agentskills.io/specification).
4
5	## Quick Start
6
7	When asked to create a new skill:
8
9	1. Gather requirements: Ask what the skill should do
10	1. Choose a name: lowercase letters, numbers, hyphens only (e.g., `pdf-processing`, `data-analysis`)
11	1. Generate the structure: Create `SKILL.md` with proper frontmatter
12	1. Add optional components: scripts, references, assets as needed
13
14	## Directory Structure
15
16	```
17	skill-name/
18	├── SKILL.md # Required: main skill file
19	├── scripts/ # Optional: executable code
20	├── references/ # Optional: additional documentation
21	└── assets/ # Optional: static resources
22	```
23
24	## SKILL.md Template
25
26	```markdown
27	---
28	name: your-skill-name
29	description: What this skill does and when to use it. Be specific and include keywords that help agents identify relevant tasks. Max 1024 characters.
30	license: MIT
31	compatibility: Optional - environment requirements if any
32	metadata:
33	author: your-name
34	version: "1.0"
35	allowed-tools: Optional - pre-approved tools (experimental)
36	---
37
38	# Skill Title
39
40	Brief introduction to the skill.
41
42	## Usage
43
44	Step-by-step instructions on how to use this skill.
45
46	## Examples
47
48	Example inputs and outputs.
49
50	## Notes
51
52	Common edge cases and tips.
53	```
54
55	## Field Requirements
56
57	### name (required)
58
59	- 1-64 characters
60	- Lowercase letters, numbers, hyphens only
61	- Cannot start or end with `-`
62	- No consecutive hyphens `--`
63	- Must match directory name
64
65	Valid: `pdf-processing`, `data-analysis`, `code-review-2`
66	Invalid: `PDF-Processing`, `-pdf`, `pdf--processing`
67
68	### description (required)
69
70	- 1-1024 characters
71	- Describe WHAT the skill does AND WHEN to use it
72	- Include specific keywords for discoverability
73
74	Good: "Extracts text and tables from PDF files. Use when working with PDF documents, extracting content from PDFs, or processing scanned documents."
75	Poor: "Helps with PDFs."
76
77	### license (optional)
78
79	- License name or reference to bundled license file
80	- Examples: `MIT`, `Apache-2.0`, `Proprietary. LICENSE.txt has complete terms`
81
82	### compatibility (optional)
83
84	- 1-500 characters
85	- Only include if skill has specific environment requirements
86	- Examples: "Requires Python 3.8+ and pandas", "Needs internet access for API calls"
87
88	### metadata (optional)
89
90	- Arbitrary key-value pairs
91	- Common keys: `author`, `version`, `tags`
92
93	### allowed-tools (optional, experimental)
94
95	- Space-delimited list of pre-approved tools
96	- Example: `Bash(git:) Bash(jq:) Read`
97
98	## Best Practices
99
100	### Progressive Disclosure
101
102	Design for efficient context usage:
103
104	1. Metadata (~100 tokens): Loaded at startup for all skills
105	1. Instructions (\<5000 tokens recommended): Loaded when skill is activated
106	1. Resources: Loaded on-demand
107
108	Keep `SKILL.md` under 500 lines. Move detailed content to `references/`.
109
110	### File Organization
111
112	- Keep `SKILL.md` focused on core instructions
113	- Put detailed docs in `references/REFERENCE.md`
114	- Put templates in `assets/`
115	- Put executable code in `scripts/`
116
117	### File References
118
119	Use relative paths from skill root:
120
121	```markdown
122	See [the reference guide](references/REFERENCE.md) for details.
123	Run: scripts/process.py
124	```
125
126	Keep references one level deep. Avoid deeply nested chains.
127
128	## Validation
129
130	After creating a skill, validate it:
131
132	```bash
133	# Install skills-ref if needed
134	npm install -g @agentskills/skills-ref
135
136	# Validate the skill
137	skills-ref validate ./your-skill-name
138	```
139
140	## Workflow Example
141
142	When asked to create a skill for X:
143
144	1. Create directory: `your-skill-name/`
145	1. Write `SKILL.md` with:
146	- Proper frontmatter (name, description)
147	- Clear instructions in Markdown body
148	1. Optionally add:
149	- `scripts/` for helper scripts
150	- `references/` for detailed docs
151	- `assets/` for templates/data
152	1. Validate with `skills-ref validate`
153	1. Test the skill with an agent
154
155	## Common Patterns
156
157	### Simple Skill
158
159	Just a `SKILL.md` with instructions:
160
161	```
162	my-skill/
163	└── SKILL.md
164	```
165
166	### Skill with Scripts
167
168	For skills that run code:
169
170	```
171	my-skill/
172	├── SKILL.md
173	└── scripts/
174	└── helper.py
175	```
176
177	### Skill with References
178
179	For detailed documentation:
180
181	```
182	my-skill/
183	├── SKILL.md
184	└── references/
185	├── REFERENCE.md
186	└── examples.md
187	```
188
189	### Full-featured Skill
190
191	```
192	my-skill/
193	├── SKILL.md
194	├── scripts/
195	│ ├── setup.sh
196	│ └── process.py
197	├── references/
198	│ ├── API.md
199	│ └── FORMATS.md
200	└── assets/
201	├── template.json
202	└── schema.json
203	```
204
205	## References
206
207	- [Official Specification](https://agentskills.io/specification)
208	- [Documentation Index](https://agentskills.io/llms.txt)
209	- [skills-ref Validator](https://githu

People who liked this also viewed

6 picks

Type

create-skill

113stars23forks

View on GitHub

How to install create-skill?

jinzhezenggroup/computational-chemistry-agent-skills/create-skill

$npx -y skills add jinzhezenggroup/computational-chemistry-agent-skills --skill create-skill

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Create Skill
2
3	This skill helps you create new Agent Skills that follow the [agentskills.io specification](https://agentskills.io/specification).
4
5	## Quick Start
6
7	When asked to create a new skill:
8
9	1. Gather requirements: Ask what the skill should do
10	1. Choose a name: lowercase letters, numbers, hyphens only (e.g., `pdf-processing`, `data-analysis`)
11	1. Generate the structure: Create `SKILL.md` with proper frontmatter
12	1. Add optional components: scripts, references, assets as needed
13
14	## Directory Structure
15
16	```
17	skill-name/
18	├── SKILL.md # Required: main skill file
19	├── scripts/ # Optional: executable code
20	├── references/ # Optional: additional documentation
21	└── assets/ # Optional: static resources
22	```
23
24	## SKILL.md Template
25
26	```markdown
27	---
28	name: your-skill-name
29	description: What this skill does and when to use it. Be specific and include keywords that help agents identify relevant tasks. Max 1024 characters.
30	license: MIT
31	compatibility: Optional - environment requirements if any
32	metadata:
33	author: your-name
34	version: "1.0"
35	allowed-tools: Optional - pre-approved tools (experimental)
36	---
37
38	# Skill Title
39
40	Brief introduction to the skill.
41
42	## Usage
43
44	Step-by-step instructions on how to use this skill.
45
46	## Examples
47
48	Example inputs and outputs.
49
50	## Notes
51
52	Common edge cases and tips.
53	```
54
55	## Field Requirements
56
57	### name (required)
58
59	- 1-64 characters
60	- Lowercase letters, numbers, hyphens only
61	- Cannot start or end with `-`
62	- No consecutive hyphens `--`
63	- Must match directory name
64
65	Valid: `pdf-processing`, `data-analysis`, `code-review-2`
66	Invalid: `PDF-Processing`, `-pdf`, `pdf--processing`
67
68	### description (required)
69
70	- 1-1024 characters
71	- Describe WHAT the skill does AND WHEN to use it
72	- Include specific keywords for discoverability
73
74	Good: "Extracts text and tables from PDF files. Use when working with PDF documents, extracting content from PDFs, or processing scanned documents."
75	Poor: "Helps with PDFs."
76
77	### license (optional)
78
79	- License name or reference to bundled license file
80	- Examples: `MIT`, `Apache-2.0`, `Proprietary. LICENSE.txt has complete terms`
81
82	### compatibility (optional)
83
84	- 1-500 characters
85	- Only include if skill has specific environment requirements
86	- Examples: "Requires Python 3.8+ and pandas", "Needs internet access for API calls"
87
88	### metadata (optional)
89
90	- Arbitrary key-value pairs
91	- Common keys: `author`, `version`, `tags`
92
93	### allowed-tools (optional, experimental)
94
95	- Space-delimited list of pre-approved tools
96	- Example: `Bash(git:) Bash(jq:) Read`
97
98	## Best Practices
99
100	### Progressive Disclosure
101
102	Design for efficient context usage:
103
104	1. Metadata (~100 tokens): Loaded at startup for all skills
105	1. Instructions (\<5000 tokens recommended): Loaded when skill is activated
106	1. Resources: Loaded on-demand
107
108	Keep `SKILL.md` under 500 lines. Move detailed content to `references/`.
109
110	### File Organization
111
112	- Keep `SKILL.md` focused on core instructions
113	- Put detailed docs in `references/REFERENCE.md`
114	- Put templates in `assets/`
115	- Put executable code in `scripts/`
116
117	### File References
118
119	Use relative paths from skill root:
120
121	```markdown
122	See [the reference guide](references/REFERENCE.md) for details.
123	Run: scripts/process.py
124	```
125
126	Keep references one level deep. Avoid deeply nested chains.
127
128	## Validation
129
130	After creating a skill, validate it:
131
132	```bash
133	# Install skills-ref if needed
134	npm install -g @agentskills/skills-ref
135
136	# Validate the skill
137	skills-ref validate ./your-skill-name
138	```
139
140	## Workflow Example
141
142	When asked to create a skill for X:
143
144	1. Create directory: `your-skill-name/`
145	1. Write `SKILL.md` with:
146	- Proper frontmatter (name, description)
147	- Clear instructions in Markdown body
148	1. Optionally add:
149	- `scripts/` for helper scripts
150	- `references/` for detailed docs
151	- `assets/` for templates/data
152	1. Validate with `skills-ref validate`
153	1. Test the skill with an agent
154
155	## Common Patterns
156
157	### Simple Skill
158
159	Just a `SKILL.md` with instructions:
160
161	```
162	my-skill/
163	└── SKILL.md
164	```
165
166	### Skill with Scripts
167
168	For skills that run code:
169
170	```
171	my-skill/
172	├── SKILL.md
173	└── scripts/
174	└── helper.py
175	```
176
177	### Skill with References
178
179	For detailed documentation:
180
181	```
182	my-skill/
183	├── SKILL.md
184	└── references/
185	├── REFERENCE.md
186	└── examples.md
187	```
188
189	### Full-featured Skill
190
191	```
192	my-skill/
193	├── SKILL.md
194	├── scripts/
195	│ ├── setup.sh
196	│ └── process.py
197	├── references/
198	│ ├── API.md
199	│ └── FORMATS.md
200	└── assets/
201	├── template.json
202	└── schema.json
203	```
204
205	## References
206
207	- [Official Specification](https://agentskills.io/specification)
208	- [Documentation Index](https://agentskills.io/llms.txt)
209	- [skills-ref Validator](https://githu

People who liked this also viewed

6 picks

Type