source-driven-development

Name: source-driven-development
Author: jmxt3

32stars3forks

How to install source-driven-development?

jmxt3/gitscape.ai/source-driven-development

$npx -y skills add jmxt3/gitscape.ai --skill source-driven-development

Installs into the current project.

Summary

Grounds every implementation decision in official documentation. Use when you want authoritative, source-cited code free from outdated patterns. Use when building with any framework or library where correctness matters.

Files · 1

View on GitHub

SKILL.md

1	# Source-Driven Development
2
3	## Overview
4
5	Every framework decision and library usage should be grounded in official documentation. AI agents have training cutoffs and can produce outdated, deprecated, or hallucinated API patterns. Source-driven development means: verify before you code, cite what you verified, and flag what you couldn't confirm.
6
7	## When to Use
8
9	- Using any external library or framework
10	- Implementing a pattern that isn't already established in the codebase
11	- Unsure whether an API or approach is current
12	- Seeing a pattern that looks right but isn't confirmed in the existing codebase
13
14	## The Process
15
16	### Step 1: Identify What Needs Verification
17
18	Before writing any code that uses an external library:
19
20	1. List the specific APIs, functions, or patterns you plan to use
21	2. Note the version you're targeting (check `requirements.txt`, `package.json`, or `pyproject.toml`)
22	3. Flag anything you're not 100% confident is current
23
24	### Step 2: Verify in Official Sources
25
26	Priority order for sources:
27	1. Official docs — the library's own documentation site
28	2. Official GitHub — the library's `README.md`, `CHANGELOG.md`, or `examples/`
29	3. Release notes — for the specific version in use
30	4. Existing codebase — how the library is already used in this project
31
32	Never use as authoritative sources:
33	- Stack Overflow answers (can be outdated)
34	- Blog posts (may target a different version)
35	- Your own training knowledge without verification
36	- ChatGPT or other AI outputs as a source of truth
37
38	### Step 3: Cite Your Sources
39
40	When writing code that uses verified patterns, leave a reference comment for future maintainers:
41
42	```python
43	# FastAPI dependency injection pattern
44	# Source: https://fastapi.tiangolo.com/tutorial/dependencies/
45	async def get_db() -> AsyncGenerator[AsyncSession, None]:
46	async with AsyncSessionLocal() as session:
47	yield session
48	```
49
50	```typescript
51	// React Query v5 data fetching
52	// Source: https://tanstack.com/query/v5/docs/framework/react/guides/queries
53	const { data, isLoading } = useQuery({
54	queryKey: ['skills', repoId],
55	queryFn: () => fetchSkills(repoId),
56	});
57	```
58
59	### Step 4: Flag What You Couldn't Verify
60
61	If you're using a pattern you cannot confirm is current, say so explicitly:
62
63	```python
64	# NOTE: Verify this middleware pattern against the current FastAPI docs.
65	# Used based on prior codebase patterns — may have changed in recent versions.
66	app.add_middleware(GZipMiddleware, minimum_size=1000)
67	```
68
69	And note it in the PR description or task update.
70
71	## GitScape-Specific Sources
72
73	For the GitScape stack, always verify against:
74
75	\| Technology \| Official Source \|
76	\|---\|---\|
77	\| FastAPI \| https://fastapi.tiangolo.com/ \|
78	\| Google Gen AI SDK (Python) \| https://googleapis.github.io/python-genai/ \|
79	\| Google Cloud Run \| https://cloud.google.com/run/docs \|
80	\| GitHub API (PyGithub) \| https://pygithub.readthedocs.io/ \|
81	\| React / TypeScript \| https://react.dev/ \|
82	\| Vite \| https://vitejs.dev/ \|
83	\| Cloud Build \| https://cloud.google.com/build/docs \|
84
85	Check the version in `api/requirements.txt` and `web/package.json` before citing docs for a specific version.
86
87	## Anti-Patterns to Avoid
88
89	\| Anti-Pattern \| Why It Fails \|
90	\|---\|---\|
91	\| Using an API without checking the version \| Libraries break across major versions \|
92	\| Copying a pattern from memory \| Training data may lag 12-24 months behind current releases \|
93	\| "This looks right" as verification \| Looks right ≠ is right \|
94	\| Citing an AI as a source \| AI outputs require human-verified sources, not circular citation \|
95
96	## Common Rationalizations
97
98	\| Rationalization \| Reality \|
99	\|---\|---\|
100	\| "I know this library well" \| APIs change. Verify. The minute spent checking prevents the hour debugging. \|
101	\| "I'll verify later" \| You won't. And by then the wrong pattern is baked in. \|
102	\| "The existing code uses this pattern" \| Great — that's evidence. Check if the existing usage is from a current dependency version. \|
103
104	## Red Flags
105
106	- Library usage with no reference to a source
107	- Patterns that differ from every other usage in the codebase without explanation
108	- Using a deprecated API (check for deprecation warnings in docs)
109	- Citing a blog post or Stack Overflow answer as the source
110
111	## Verification
112
113	After writing code that uses external libraries:
114
115	- [ ] Every external API usage is confirmed in official docs or the existing codebase
116	- [ ] The docs version matches the version in `requirements.txt` / `package.json`
117	- [ ] Unverified patterns are explicitly flagged with a `# NOTE: verify` comment
118	- [ ] No deprecated APIs are used (or if used, the deprecation is documented)

People who liked this also viewed

6 picks

Type

source-driven-development

32stars3forks

View on GitHub

How to install source-driven-development?

jmxt3/gitscape.ai/source-driven-development

$npx -y skills add jmxt3/gitscape.ai --skill source-driven-development

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Source-Driven Development
2
3	## Overview
4
5	Every framework decision and library usage should be grounded in official documentation. AI agents have training cutoffs and can produce outdated, deprecated, or hallucinated API patterns. Source-driven development means: verify before you code, cite what you verified, and flag what you couldn't confirm.
6
7	## When to Use
8
9	- Using any external library or framework
10	- Implementing a pattern that isn't already established in the codebase
11	- Unsure whether an API or approach is current
12	- Seeing a pattern that looks right but isn't confirmed in the existing codebase
13
14	## The Process
15
16	### Step 1: Identify What Needs Verification
17
18	Before writing any code that uses an external library:
19
20	1. List the specific APIs, functions, or patterns you plan to use
21	2. Note the version you're targeting (check `requirements.txt`, `package.json`, or `pyproject.toml`)
22	3. Flag anything you're not 100% confident is current
23
24	### Step 2: Verify in Official Sources
25
26	Priority order for sources:
27	1. Official docs — the library's own documentation site
28	2. Official GitHub — the library's `README.md`, `CHANGELOG.md`, or `examples/`
29	3. Release notes — for the specific version in use
30	4. Existing codebase — how the library is already used in this project
31
32	Never use as authoritative sources:
33	- Stack Overflow answers (can be outdated)
34	- Blog posts (may target a different version)
35	- Your own training knowledge without verification
36	- ChatGPT or other AI outputs as a source of truth
37
38	### Step 3: Cite Your Sources
39
40	When writing code that uses verified patterns, leave a reference comment for future maintainers:
41
42	```python
43	# FastAPI dependency injection pattern
44	# Source: https://fastapi.tiangolo.com/tutorial/dependencies/
45	async def get_db() -> AsyncGenerator[AsyncSession, None]:
46	async with AsyncSessionLocal() as session:
47	yield session
48	```
49
50	```typescript
51	// React Query v5 data fetching
52	// Source: https://tanstack.com/query/v5/docs/framework/react/guides/queries
53	const { data, isLoading } = useQuery({
54	queryKey: ['skills', repoId],
55	queryFn: () => fetchSkills(repoId),
56	});
57	```
58
59	### Step 4: Flag What You Couldn't Verify
60
61	If you're using a pattern you cannot confirm is current, say so explicitly:
62
63	```python
64	# NOTE: Verify this middleware pattern against the current FastAPI docs.
65	# Used based on prior codebase patterns — may have changed in recent versions.
66	app.add_middleware(GZipMiddleware, minimum_size=1000)
67	```
68
69	And note it in the PR description or task update.
70
71	## GitScape-Specific Sources
72
73	For the GitScape stack, always verify against:
74
75	\| Technology \| Official Source \|
76	\|---\|---\|
77	\| FastAPI \| https://fastapi.tiangolo.com/ \|
78	\| Google Gen AI SDK (Python) \| https://googleapis.github.io/python-genai/ \|
79	\| Google Cloud Run \| https://cloud.google.com/run/docs \|
80	\| GitHub API (PyGithub) \| https://pygithub.readthedocs.io/ \|
81	\| React / TypeScript \| https://react.dev/ \|
82	\| Vite \| https://vitejs.dev/ \|
83	\| Cloud Build \| https://cloud.google.com/build/docs \|
84
85	Check the version in `api/requirements.txt` and `web/package.json` before citing docs for a specific version.
86
87	## Anti-Patterns to Avoid
88
89	\| Anti-Pattern \| Why It Fails \|
90	\|---\|---\|
91	\| Using an API without checking the version \| Libraries break across major versions \|
92	\| Copying a pattern from memory \| Training data may lag 12-24 months behind current releases \|
93	\| "This looks right" as verification \| Looks right ≠ is right \|
94	\| Citing an AI as a source \| AI outputs require human-verified sources, not circular citation \|
95
96	## Common Rationalizations
97
98	\| Rationalization \| Reality \|
99	\|---\|---\|
100	\| "I know this library well" \| APIs change. Verify. The minute spent checking prevents the hour debugging. \|
101	\| "I'll verify later" \| You won't. And by then the wrong pattern is baked in. \|
102	\| "The existing code uses this pattern" \| Great — that's evidence. Check if the existing usage is from a current dependency version. \|
103
104	## Red Flags
105
106	- Library usage with no reference to a source
107	- Patterns that differ from every other usage in the codebase without explanation
108	- Using a deprecated API (check for deprecation warnings in docs)
109	- Citing a blog post or Stack Overflow answer as the source
110
111	## Verification
112
113	After writing code that uses external libraries:
114
115	- [ ] Every external API usage is confirmed in official docs or the existing codebase
116	- [ ] The docs version matches the version in `requirements.txt` / `package.json`
117	- [ ] Unverified patterns are explicitly flagged with a `# NOTE: verify` comment
118	- [ ] No deprecated APIs are used (or if used, the deprecation is documented)

People who liked this also viewed

6 picks

Type