generate-skill

Name: generate-skill
Author: TanStack

314stars19forks

How to install generate-skill?

tanstack/intent/generate-skill

$npx -y skills add TanStack/intent --skill generate-skill

Installs into the current project.

Summary

Generate a complete SKILL.md file for a library from source documentation and skill tree artifacts. Activate when bootstrapping skills for a new library, regenerating a stale skill after source changes, or producing a skill from a skill_tree.yaml entry. Takes a skill name, descri

Files · 1

View on GitHub

SKILL.md

1	# Skill Generation
2
3	You are generating a SKILL.md file for the `@tanstack/intent` agent skills
4	repo. Skills in this repo are written for coding agents (Claude Code, Cursor,
5	Copilot, Warp Oz, Codex), not for human readers. Your output will be loaded
6	into an agent's context window and used to guide code generation.
7
8	There are two modes. Detect which applies.
9
10	Mode A — Generate from domain map: A `domain_map.yaml` and `skill_spec.md`
11	exist. Generate the skill specified by name from these artifacts plus the
12	source documentation they reference.
13
14	Mode B — Generate from raw docs: No domain map exists. Generate directly
15	from source documentation provided as input.
16
17	---
18
19	## Inputs
20
21	You will receive:
22
23	If the maintainer uses a custom skills root, replace `skills/` in any paths
24	below with their chosen directory.
25
26	Monorepo: When the skill tree entry has a `package` field, write the
27	SKILL.md into that package's skills directory (e.g.
28	`packages/client/skills/core/SKILL.md`), not a shared root.
29
30	1. Skill path — format `library-group/skill-name` (e.g. `tanstack-query/core`,
31	`tanstack-router/loaders`, `db/core/live-queries`). This is the skill's
32	directory path, and it is where the `SKILL.md` lives. Only its **last
33	segment** becomes the frontmatter `name` (a spec-legal leaf): the full path
34	is never written into `name`, and `name` must contain no slashes.
35	2. Skill description — what the skill covers and when an agent should load it
36	3. Source documentation — the docs, guides, API references, and/or source
37	files to distill from
38	4. Domain map entry (Mode A only) — the skill's entry from `domain_map.yaml`
39	including failure modes, subsystems, compositions, and source references
40
41	---
42
43	## Step 1 — Determine skill type
44
45	Read the inputs and classify the skill type:
46
47	\| Type \| When to use \|
48	\| ------------- \| ---------------------------------------------------------- \|
49	\| `core` \| Framework-agnostic concepts, configuration, patterns \|
50	\| `sub-skill` \| A focused sub-topic within a core or framework skill \|
51	\| `framework` \| Framework-specific bindings, hooks, components \|
52	\| `lifecycle` \| Cross-cutting developer journey (getting started, go-live) \|
53	\| `composition` \| Integration between two or more libraries \|
54	\| `security` \| Audit checklist or security validation \|
55
56	The skill type determines the frontmatter and body structure. See
57	skill-tree-generator for the full spec of each type.
58
59	---
60
61	### Subagent guidance for batch generation
62
63	When generating multiple skills, spawn a separate subagent for each skill
64	(or per-package group). Each subagent receives the domain_map.yaml,
65	skill_tree.yaml, and the source docs relevant to its skill. This prevents
66	context bleed between skills and allows parallel generation.
67
68	---
69
70	## Step 2 — Extract content from sources
71
72	Line budget: Each SKILL.md must stay under 500 lines. Before writing,
73	estimate the content size. If a skill has 5+ failure modes, 3+ primary
74	patterns, and subsystem details, proactively plan reference files during
75	extraction — don't wait until the skill exceeds the limit.
76
77	Read through the source documentation. Extract only what a coding agent
78	cannot already know:
79
80	### What to extract
81
82	- API shapes — function signatures, hook parameters, option objects,
83	return types. Use the actual TypeScript types from source.
84	- Setup patterns — minimum viable initialization code
85	- Primary patterns — the 2–4 most important usage patterns
86	- Configuration — defaults that matter, options that change behavior
87	- Failure modes — patterns that look correct but break. Prioritize:
88	- Migration-boundary mistakes (old API that agents trained on older data produce)
89	- Silent failures (no crash, wrong behavior)
90	- Framework-specific gotchas (hydration, hook rules, provider ordering)
91	- Constraints and invariants — ordering requirements, lifecycle rules,
92	things enforced by runtime assertions
93	- Issue/discussion-sourced patterns — real developer mistakes and
94	confusion surfaced from GitHub issues and discussions (see below)
95
96	### 2b — Scan GitHub issues and discussions
97
98	Bef

People who liked this also viewed

6 picks

Type

generate-skill

314stars19forks

View on GitHub

How to install generate-skill?

tanstack/intent/generate-skill

$npx -y skills add TanStack/intent --skill generate-skill

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Skill Generation
2
3	You are generating a SKILL.md file for the `@tanstack/intent` agent skills
4	repo. Skills in this repo are written for coding agents (Claude Code, Cursor,
5	Copilot, Warp Oz, Codex), not for human readers. Your output will be loaded
6	into an agent's context window and used to guide code generation.
7
8	There are two modes. Detect which applies.
9
10	Mode A — Generate from domain map: A `domain_map.yaml` and `skill_spec.md`
11	exist. Generate the skill specified by name from these artifacts plus the
12	source documentation they reference.
13
14	Mode B — Generate from raw docs: No domain map exists. Generate directly
15	from source documentation provided as input.
16
17	---
18
19	## Inputs
20
21	You will receive:
22
23	If the maintainer uses a custom skills root, replace `skills/` in any paths
24	below with their chosen directory.
25
26	Monorepo: When the skill tree entry has a `package` field, write the
27	SKILL.md into that package's skills directory (e.g.
28	`packages/client/skills/core/SKILL.md`), not a shared root.
29
30	1. Skill path — format `library-group/skill-name` (e.g. `tanstack-query/core`,
31	`tanstack-router/loaders`, `db/core/live-queries`). This is the skill's
32	directory path, and it is where the `SKILL.md` lives. Only its **last
33	segment** becomes the frontmatter `name` (a spec-legal leaf): the full path
34	is never written into `name`, and `name` must contain no slashes.
35	2. Skill description — what the skill covers and when an agent should load it
36	3. Source documentation — the docs, guides, API references, and/or source
37	files to distill from
38	4. Domain map entry (Mode A only) — the skill's entry from `domain_map.yaml`
39	including failure modes, subsystems, compositions, and source references
40
41	---
42
43	## Step 1 — Determine skill type
44
45	Read the inputs and classify the skill type:
46
47	\| Type \| When to use \|
48	\| ------------- \| ---------------------------------------------------------- \|
49	\| `core` \| Framework-agnostic concepts, configuration, patterns \|
50	\| `sub-skill` \| A focused sub-topic within a core or framework skill \|
51	\| `framework` \| Framework-specific bindings, hooks, components \|
52	\| `lifecycle` \| Cross-cutting developer journey (getting started, go-live) \|
53	\| `composition` \| Integration between two or more libraries \|
54	\| `security` \| Audit checklist or security validation \|
55
56	The skill type determines the frontmatter and body structure. See
57	skill-tree-generator for the full spec of each type.
58
59	---
60
61	### Subagent guidance for batch generation
62
63	When generating multiple skills, spawn a separate subagent for each skill
64	(or per-package group). Each subagent receives the domain_map.yaml,
65	skill_tree.yaml, and the source docs relevant to its skill. This prevents
66	context bleed between skills and allows parallel generation.
67
68	---
69
70	## Step 2 — Extract content from sources
71
72	Line budget: Each SKILL.md must stay under 500 lines. Before writing,
73	estimate the content size. If a skill has 5+ failure modes, 3+ primary
74	patterns, and subsystem details, proactively plan reference files during
75	extraction — don't wait until the skill exceeds the limit.
76
77	Read through the source documentation. Extract only what a coding agent
78	cannot already know:
79
80	### What to extract
81
82	- API shapes — function signatures, hook parameters, option objects,
83	return types. Use the actual TypeScript types from source.
84	- Setup patterns — minimum viable initialization code
85	- Primary patterns — the 2–4 most important usage patterns
86	- Configuration — defaults that matter, options that change behavior
87	- Failure modes — patterns that look correct but break. Prioritize:
88	- Migration-boundary mistakes (old API that agents trained on older data produce)
89	- Silent failures (no crash, wrong behavior)
90	- Framework-specific gotchas (hydration, hook rules, provider ordering)
91	- Constraints and invariants — ordering requirements, lifecycle rules,
92	things enforced by runtime assertions
93	- Issue/discussion-sourced patterns — real developer mistakes and
94	confusion surfaced from GitHub issues and discussions (see below)
95
96	### 2b — Scan GitHub issues and discussions
97
98	Bef

People who liked this also viewed

6 picks

Type