| name | use-index |
| description | Navigate the .5/index/ codebase index to quickly identify target files and understand project structure. Use before running broad Glob/Grep scans when you need to find where a feature lives, where new code should go, which routes exist, or what the data model looks like. |
| allowed-tools | Read, Bash |
| user-invocable | false |
Use Index
Use .5/index/ to orient in the codebase quickly before resorting to broad file scans.
Step 1: Check existence
ls .5/index/ 2>/dev/null
If the directory is missing or empty, inform the user they can run /5:reconfigure to generate it, then fall back to Glob/Grep.
Step 2: Read the manifest
Read .5/index/README.md. It lists every generated index file, its purpose, and the generation timestamp.
Step 3: Check freshness
Look for the Generated: timestamp in README.md.
- Fresh (≤ 1 day): proceed with the index.
- Stale (> 1 day): run
.5/index/rebuild-index.sh to rebuild it, then proceed with the refreshed index.
Step 4: Read only the relevant file(s)
Pick the file that matches the task. Skip unrelated ones.
| Task | File |
|---|
| Where does new code go / what modules exist? | modules.md |
| API surface, HTTP routes, or handlers | routes.md |
| Data models, schemas, or migrations | models.md |
| Shared helpers, utilities, or exported functions | libraries.md |
| Build, run, test, or dev-workflow commands | commands.md |
Step 5: Interpret entries
Each file has compact entries — path plus a short descriptor (HTTP method, exported name, key field, etc.). Use them to identify target files directly. Do not read the full index file if a handful of entries answer the question.
Step 6: Fall back gracefully
If a file is too sparse or an area is missing from the index, note the gap and fall back to a targeted Grep/Glob for that area only. Do not re-scan sections already covered by a fresh index entry.