learn

1## EXECUTE NOW
2 
3**Topic: $ARGUMENTS**
4 
5Parse immediately:
6- If topic provided: research that topic
7- If topic empty: read `self/goals.md` for highest-priority unexplored direction and propose it
8- If topic includes `--deep`/`--light`/`--moderate`: force that depth, strip flag from topic
9- If no topic and no goals.md: ask "What would you like to research?"
10 
11**Steps:**
12 
131. **Read config** — tool preferences, depth, domain vocabulary
142. **Determine depth** — from flags, config default, or fallback to moderate
153. **Research** — tool cascade: primary → fallback → last resort
164. **File to inbox** — with full provenance metadata
175. **Chain to processing** — next step based on pipeline chaining mode
186. **Update goals.md** — append new research directions discovered
19 
20**START NOW.** Reference below explains methodology.
21 
22---
23 
24## Step 1: Read Configuration
25 
26```
27ops/config.yaml             — research tools, depth, pipeline chaining
28ops/derivation-manifest.md  — domain vocabulary (inbox folder, reduce skill name)
29```
30 
31**From config.yaml** (defaults if missing):
32```yaml
33research:
34  primary: exa-deep-research      # exa-deep-research | exa-web-search | web-search
35  fallback: exa-web-search
36  last_resort: web-search
37  default_depth: moderate          # light | moderate | deep
38pipeline:
39  chaining: suggested             # manual | suggested | automatic
40```
41 
42**From derivation-manifest.md** (universal defaults if missing):
43- Inbox folder: `inbox/` (could be `journal/`, `encounters/`, etc.)
44…