skill-maker

Name: skill-maker
Author: redhat-developer

38installs31forks

How to install skill-maker?

redhat-developer/rhdh-skill/skill-maker

$npx -y skills add redhat-developer/rhdh-skill --skill skill-maker

Installs into the current project.

Summary

Create, audit, or consolidate agent skills following the Agent Skills open standard (agentskills.io). Interviews the user relentlessly about intent, scope, and edge cases before drafting. Covers SKILL.md structure, frontmatter, progressive disclosure, description optimization, sc

Files · 1

View on GitHub

SKILL.md

1	<intake>
2
3	# Create, Audit, or Consolidate Skills
4
5	Create agent skills following the [Agent Skills open standard](https://agentskills.io/specification).
6
7	What do you need to do?
8
9	1. Audit an existing skill — Review, improve, or debug a SKILL.md
10	2. Create a new skill — Interview, draft, and review from scratch
11	3. Consolidate skills — Merge multiple skills into fewer
12
13	Wait for response before proceeding.
14	</intake>
15
16	<routing>
17
18	\| Response \| Workflow \|
19	\|----------\|----------\|
20	\| 1, "audit", "review", "check", "fix", "improve" \| Audit Workflow (Step 1–4 in this file) \|
21	\| 2, "create", "write", "build", "new", "draft" \| Phases 1–5 (Interview → Draft → Description → Scripts → Review) in this file \|
22	\| 3, "consolidate", "merge", "combine" \| `references/consolidation-guide.md` — return to Phase 5 for final checklist \|
23
24	</routing>
25
26	## Audit Workflow
27
28	Use this workflow when reviewing, improving, or debugging an existing skill.
29
30	### Step 1: Locate and read the skill
31
32	Read the full SKILL.md and list all files in the skill directory (`references/`, `scripts/`, `templates/`, `assets/`).
33
34	### Step 2: Run the audit checklist
35
36	Check each category. Note issues as you go.
37
38	Frontmatter:
39
40	- [ ] `name` matches the directory name, lowercase+hyphens, max 64 chars
41	- [ ] `description` is under 1024 chars, non-empty, third person
42	- [ ] `description` includes trigger phrases (not just a summary of what the skill does)
43	- [ ] `description` covers edge phrasings users would actually say
44
45	Structure:
46
47	- [ ] SKILL.md body is under 500 lines
48	- [ ] Essential principles are inline in SKILL.md (not only in a reference file)
49	- [ ] All referenced files exist (check every path in the SKILL.md)
50	- [ ] References are one level deep (no nested chains: A → B → C)
51
52	Content quality:
53
54	- [ ] No rigid ALWAYS/NEVER rules without reasoning (explain WHY)
55	- [ ] No explanations of things the agent already knows from training
56	- [ ] Steps are specific and verifiable (not "handle errors appropriately")
57	- [ ] Success criteria are observable and testable
58	- [ ] Examples use fake data where appropriate
59
60	Router pattern (if applicable):
61
62	- [ ] Intake question asks what the user wants before routing
63	- [ ] Router table maps commands to reference files
64	- [ ] All referenced workflow/reference files exist
65	- [ ] Essential principles are in SKILL.md, not only in sub-command references
66	- [ ] If skill has multiple semantic sections, consider XML tags for structure (see `references/xml-structure-guide.md`)
67
68	Scripts (if present):
69
70	- [ ] Scripts have shebangs, `--help`, and structured output
71	- [ ] No interactive prompts (all input via flags/env/stdin)
72	- [ ] Cross-platform paths (pathlib, no hardcoded separators)
73	- [ ] Error messages explain what went wrong and what to do
74
75	Read `references/anti-patterns.md` for the full catalog of common failures.
76
77	### Step 3: Generate the report
78
79	Present findings grouped by severity:
80
81	1. Critical — skill won't trigger or produces wrong output
82	2. Important — structural issues, missing files, spec violations
83	3. Minor — style, conciseness, optimization opportunities
84
85	For each finding, state the issue, cite the specific line or section, and recommend a fix.
86
87	### Step 4: Offer fixes
88
89	Ask the user which findings to fix. Apply changes surgically — don't rewrite sections that aren't broken. Run the Phase 5 review checklist on the modified skill before finishing.
90
91	## Phase 1: Interview
92
93	Interview the user about every aspect of this skill until reaching shared understanding. Walk down each branch of the design tree, resolving dependencies between decisions one-by-one. For each question, provide your recommended answer.
94
95	### Interview cadence
96
97	Ask one question at a time. Wait for the answer before asking the next. Adapt follow-ups based on what you learn. Each question should provide clear benefit toward building a better skill — cut questions the codebase can answer for you.
98
99	If a question can be answered by exploring the codebase, explore the codebase instead of asking.
100
101	Focus areas, roughly in order:
102
103	1. Purpose and audience.

People who liked this also viewed

6 picks

Type

skill-maker

38installs31forks

View on GitHub

How to install skill-maker?

redhat-developer/rhdh-skill/skill-maker

$npx -y skills add redhat-developer/rhdh-skill --skill skill-maker

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	<intake>
2
3	# Create, Audit, or Consolidate Skills
4
5	Create agent skills following the [Agent Skills open standard](https://agentskills.io/specification).
6
7	What do you need to do?
8
9	1. Audit an existing skill — Review, improve, or debug a SKILL.md
10	2. Create a new skill — Interview, draft, and review from scratch
11	3. Consolidate skills — Merge multiple skills into fewer
12
13	Wait for response before proceeding.
14	</intake>
15
16	<routing>
17
18	\| Response \| Workflow \|
19	\|----------\|----------\|
20	\| 1, "audit", "review", "check", "fix", "improve" \| Audit Workflow (Step 1–4 in this file) \|
21	\| 2, "create", "write", "build", "new", "draft" \| Phases 1–5 (Interview → Draft → Description → Scripts → Review) in this file \|
22	\| 3, "consolidate", "merge", "combine" \| `references/consolidation-guide.md` — return to Phase 5 for final checklist \|
23
24	</routing>
25
26	## Audit Workflow
27
28	Use this workflow when reviewing, improving, or debugging an existing skill.
29
30	### Step 1: Locate and read the skill
31
32	Read the full SKILL.md and list all files in the skill directory (`references/`, `scripts/`, `templates/`, `assets/`).
33
34	### Step 2: Run the audit checklist
35
36	Check each category. Note issues as you go.
37
38	Frontmatter:
39
40	- [ ] `name` matches the directory name, lowercase+hyphens, max 64 chars
41	- [ ] `description` is under 1024 chars, non-empty, third person
42	- [ ] `description` includes trigger phrases (not just a summary of what the skill does)
43	- [ ] `description` covers edge phrasings users would actually say
44
45	Structure:
46
47	- [ ] SKILL.md body is under 500 lines
48	- [ ] Essential principles are inline in SKILL.md (not only in a reference file)
49	- [ ] All referenced files exist (check every path in the SKILL.md)
50	- [ ] References are one level deep (no nested chains: A → B → C)
51
52	Content quality:
53
54	- [ ] No rigid ALWAYS/NEVER rules without reasoning (explain WHY)
55	- [ ] No explanations of things the agent already knows from training
56	- [ ] Steps are specific and verifiable (not "handle errors appropriately")
57	- [ ] Success criteria are observable and testable
58	- [ ] Examples use fake data where appropriate
59
60	Router pattern (if applicable):
61
62	- [ ] Intake question asks what the user wants before routing
63	- [ ] Router table maps commands to reference files
64	- [ ] All referenced workflow/reference files exist
65	- [ ] Essential principles are in SKILL.md, not only in sub-command references
66	- [ ] If skill has multiple semantic sections, consider XML tags for structure (see `references/xml-structure-guide.md`)
67
68	Scripts (if present):
69
70	- [ ] Scripts have shebangs, `--help`, and structured output
71	- [ ] No interactive prompts (all input via flags/env/stdin)
72	- [ ] Cross-platform paths (pathlib, no hardcoded separators)
73	- [ ] Error messages explain what went wrong and what to do
74
75	Read `references/anti-patterns.md` for the full catalog of common failures.
76
77	### Step 3: Generate the report
78
79	Present findings grouped by severity:
80
81	1. Critical — skill won't trigger or produces wrong output
82	2. Important — structural issues, missing files, spec violations
83	3. Minor — style, conciseness, optimization opportunities
84
85	For each finding, state the issue, cite the specific line or section, and recommend a fix.
86
87	### Step 4: Offer fixes
88
89	Ask the user which findings to fix. Apply changes surgically — don't rewrite sections that aren't broken. Run the Phase 5 review checklist on the modified skill before finishing.
90
91	## Phase 1: Interview
92
93	Interview the user about every aspect of this skill until reaching shared understanding. Walk down each branch of the design tree, resolving dependencies between decisions one-by-one. For each question, provide your recommended answer.
94
95	### Interview cadence
96
97	Ask one question at a time. Wait for the answer before asking the next. Adapt follow-ups based on what you learn. Each question should provide clear benefit toward building a better skill — cut questions the codebase can answer for you.
98
99	If a question can be answered by exploring the codebase, explore the codebase instead of asking.
100
101	Focus areas, roughly in order:
102
103	1. Purpose and audience.

People who liked this also viewed

6 picks

Type