---
name: feed
description: "Personal intelligent feed aggregation — fetch, compile knowledge base, and generate daily digests"
---

# /feed — AutoWiki Daily Pipeline

You are running the AutoWiki system. This is the orchestrator that chains the fetch and kb layers together.

## Available Commands

### `/feed` (no args) — Full pipeline: fetch → compile → digest

This runs the complete daily pipeline in sequence:

**Step 1: Fetch** (invoke /fetch)
```bash
uv run autowiki fetch
```
Wait for completion. Check output:
```bash
ls raw/YYYY-MM-DD/ | wc -l
```

**Step 2: Compile** (invoke /kb compile)
Read raw data, user profile, and interests. Then compile to kb/ following the /kb skill instructions for Phase 2.

**Step 3: Digest** (invoke /kb digest)
Generate daily digest to `kb/daily/YYYY-MM-DD.md` following the /kb skill instructions.

**Step 4: Commit**
```bash
git add raw/ kb/
git commit -m "digest: YYYY-MM-DD daily feed"
```

**Step 5: Summary**
Print a brief terminal summary of the top items.

Replace YYYY-MM-DD with today's actual date.

### `/feed add <URL>` — Add a new source

Delegates to /fetch: add the URL to feeds.yaml.

### `/feed sources` — List current sources

Delegates to /fetch: show all sources.

### `/feed explore` — Discover new sources

Delegates to /fetch: suggest new sources.

### `/feed profile` — View/edit profile

1. List all files in `profiles/`
2. Ask user what they want to change
3. Edit the relevant profile file

### `/feed search <query>` — Search knowledge base

Delegates to /kb: search across kb/.

### `/feed lint` — Health check

Delegates to /kb: run kb health check.

## Important Notes

- Always use `uv run` to execute Python commands
- The fetch layer writes to `raw/`, the kb layer reads from `raw/` and writes to `kb/` (including `kb/daily/`)
- Replace YYYY-MM-DD with today's actual date
- The entire project root is an Obsidian vault