update SKILL.md and spec-driven-development.md to implement a sequential ID system for spec folders and enhance the INDEX.md overview with formatting rules and status tracking

Author: hendrikebbers

claude-project-base/conventions/spec-driven-development.md

@@ -15,21 +15,45 @@ Specs live in a `specs/` directory in the project root. Each spec gets its own s
 
 ```
 specs/
-├── 42-user-auth-flow/
+├── INDEX.md
+├── 001-user-auth-flow/
 │   ├── design.md
 │   ├── behaviors.md
 │   └── steps.md          (optional)
-├── 87-csv-export-api/
+├── 002-csv-export-api/
 │   ├── design.md
 │   └── behaviors.md
 └── ...
 ```
 
 ### Folder Naming
 
-- **Based on a GitHub issue:** `<issue-number>-short-description` (e.g., `42-user-auth-flow`)
-- **Without an issue:** `short-description` (e.g., `csv-export-api`)
-- Keep the description to 3–4 words in kebab-case
+- Format: `<sequential-number>-short-description` (e.g., `001-user-auth-flow`, `002-csv-export-api`)
+- The sequential number is zero-padded to 3 digits and incremented from the last entry in `INDEX.md`.
+- Keep the description to 3–4 words in kebab-case.
+
+### `INDEX.md` — Spec Overview
+
+The file `specs/INDEX.md` is a central index of all specs. It provides Claude and developers with an at-a-glance overview of which specs exist, what they cover, and whether they have been implemented.
+
+Format:
+
+```markdown
+# Spec Index
+
+| ID  | Name | Description | GitHub Issue | Status |
+|-----|------|-------------|--------------|--------|
+| 001 | User auth flow | JWT-based login and registration with refresh tokens | #42 | done |
+| 002 | CSV export API | REST endpoint to export filtered datasets as CSV | #87 | open |
+| 003 | Rate limiting | Token-bucket rate limiting for all public API endpoints | — | in progress |
+```
+
+Rules:
+- Every spec must have an entry in `INDEX.md`. A spec without an index entry is incomplete.
+- **Status** values: `open` (not started), `in progress` (being implemented), `done` (implemented and verified).
+- **GitHub Issue** column contains the issue reference (e.g., `#42`) or `—` if no issue exists.
+- The index is updated whenever a spec is created or its status changes.
+- The sequential ID in the index determines the canonical order of specs.
 
 ## Files
 

claude-project-base/skills/spec-create/SKILL.md

@@ -51,7 +51,9 @@ Present the drafted issues to the user for review and adjustment. **Do not creat
 
 ### 3. Create the spec folder
 
-Create the spec folder under `specs/` following the naming convention from the spec-driven development doc.
+Determine the next sequential ID by reading `specs/INDEX.md`. If the file does not exist yet, create it with the table header (see spec-driven development doc) and start at `001`. Otherwise, increment from the highest existing ID.
+
+Create the spec folder under `specs/` using the format `<ID>-short-description` (e.g., `001-user-auth-flow`).
 
 ### 4. Write `design.md` — Interactive planning
 
@@ -96,8 +98,11 @@ Review the scenarios with the user. Ask explicitly: "Are there edge cases or err
 
 ### 6. Summary
 
-After both files are written, provide a short summary:
-- Link to the created spec folder
-- Count of behavioral scenarios
-- Any open questions from the design
-- Suggest next steps: implement manually, use `/spec-implement` for a step-by-step plan, or use `/spec-review` after implementation to verify completeness
+After both files are written:
+
+1. **Update `specs/INDEX.md`** — Add a new row with the spec's ID, name, GitHub issue reference (or `—`), and status `open`.
+2. Provide a short summary:
+   - Link to the created spec folder
+   - Count of behavioral scenarios
+   - Any open questions from the design
+   - Suggest next steps: implement manually, use `/spec-implement` for a step-by-step plan, or use `/spec-review` after implementation to verify completeness