| Dimension | Shared | Only in A | Only in B | Overlap |
|---|---|---|---|---|
| Sections | 0 | 2 | 1 | 0% |
| Commands | 0 | 0 | 3 | 0% |
| Section tags | 1 | 2 | 3 | 17% |
What each file covers
Sections
0 shared · 2 only in A · 1 only in B- − Supabase Instructions
- − Rules
- + AI Service Instructions
Commands
0 shared · 0 only in A · 3 only in B- + python -m mypy app
- + python -m ruff check .
- + python -m pytest
Section tags
1 shared · 2 only in A · 3 only in B- − code-style
- − do-not
- + test
- + lint-format
- + api
- agent-behaviour
Line diff
Aledon8/OpenLeukemia · supabase/CLAUDE.md
@@ −1 @@
1# Supabase Instructions
2
3Supabase is the MVP platform layer for auth, Postgres, storage, Row Level Security, consent records, and lightweight Edge Functions.
4
5## Rules
6
7- Treat Row Level Security as a primary protection layer.
8- Keep consent behavior explicit, auditable, and revocable.
9- Prefer Supabase for routine product workflows and `ai-service/` for Python, ML, OCR, extraction, and compute-heavy work.
10- Keep migrations and policies readable and reversible where practical.
11- Do not weaken privacy, authorization, or consent checks without an explicit product reason.
12- For schema or policy changes, plan affected data flows and access paths before editing.
13
Aledon8/OpenLeukemia · .github/instructions/ai-service.instructions.md
@@ +1 @@
1---
2applyTo: "ai-service/**/*.py"
3---
4
5# AI Service Instructions
6
7- Keep FastAPI route handlers thin; put reusable behavior in `app/domains`.
8- Use explicit Pydantic request and response models.
9- Preserve strict typing and keep `python -m mypy app` clean.
10- Use `response_model` for API route outputs.
11- Treat extraction and model outputs as explainable signals or candidate data, not clinical decisions.
12- Run `python -m ruff check .` and the relevant `python -m pytest` tests for Python changes.
13
@@ −1 +1 @@
1−# Supabase Instructions
1+---
2+applyTo: "ai-service/**/*.py"
3+---
24
3−Supabase is the MVP platform layer for auth, Postgres, storage, Row Level Security, consent records, and lightweight Edge Functions.
5+# AI Service Instructions
46
5−## Rules
6−
7−- Treat Row Level Security as a primary protection layer.
8−- Keep consent behavior explicit, auditable, and revocable.
9−- Prefer Supabase for routine product workflows and `ai-service/` for Python, ML, OCR, extraction, and compute-heavy work.
10−- Keep migrations and policies readable and reversible where practical.
11−- Do not weaken privacy, authorization, or consent checks without an explicit product reason.
12−- For schema or policy changes, plan affected data flows and access paths before editing.
7+- Keep FastAPI route handlers thin; put reusable behavior in `app/domains`.
8+- Use explicit Pydantic request and response models.
9+- Preserve strict typing and keep `python -m mypy app` clean.
10+- Use `response_model` for API route outputs.
11+- Treat extraction and model outputs as explainable signals or candidate data, not clinical decisions.
12+- Run `python -m ruff check .` and the relevant `python -m pytest` tests for Python changes.
1313
