---
name: path
description: Compile a free-text objective (e.g. "FAANG interview prep") into a pathway overlay the graph UI renders — an ordered route through topics with per-segment strength and study suggestions. Use when the user states a goal that isn't (or spans) predefined goals.
---

# /path — compile an objective into a pathway overlay

**First-touch (run this first):** `.venv/bin/python -m brain first-touch path`. If it
prints a paragraph, this is the user's first pathway — open your reply with that text
verbatim, then continue. If it prints nothing, skip it and proceed. (No source flag to
set here: the explainer retires once the first `ui/paths/<slug>.json` overlay is written.)

1. Understand the objective; find what it touches:
   `.venv/bin/python -m brain search "<objective key phrases>" -k 8`, read
   `goals/goals.yaml` + relevant `goals/roadmaps/*.yaml`, and `brain gaps` output.
2. Select the route: the topics (graph node ids — knowledge topics where evidenced,
   roadmap topic ids where missing) that genuinely serve the objective, ordered by
   prerequisite structure then by leverage. 6–14 nodes; a route, not an inventory.
3. Mark each node's status from evidence: `strong` (evidenced ≥3), `weak` (1–2.9),
   `missing` (0). Use `brain graph` + `ui/graph.json` for current numbers — never
   guess.
4. Write `ui/paths/<slug>.json`:
   ```json
   {
     "title": "FAANG interview prep",
     "objective": "<one sentence>",
     "created": "<today>",
     "nodes": [{"id": "big-o", "status": "missing"}, ...],
     "suggestions": [
       {"text": "<specific ordered action for a weak/missing segment>", "nodeId": "big-o"},
       ...
     ]
   }
   ```
   Suggestions: 3–6, concrete and sequenced ("do X before Y because Z"), tied to
   route segments via nodeId. Ground every claim about his knowledge in the data.
5. Overlays are commit-worthy artifacts; stale ones can be regenerated by
   re-running /path with the same objective.

**Finish every run (ux.md #2/#3/#6 — one command per tool call):**
1. `.venv/bin/python -m brain graph` (bundles the new overlay; no notes were
   written, so no ingest needed).
2. Snapshot: `git add` the overlay file, then
   `git commit -m "snapshot: path: <objective slug>"`. No git? One line:
   snapshot skipped, pathway still saved.
3. End with the receipt block (docs/ux.md #2): overlay file + node count and
   status mix (strong/weak/missing); "map data refreshed — open brain ui and
   pick the pathway"; "saved a local snapshot — nothing leaves your machine";
   next action (usually the first suggestion on the route).