session-search

1# Session Search
2 
3Answers "which of my past sessions was about X" and then pulls that session's context in.
4 
5## Step 1: Check the graph exists and is fresh
6 
7```bash
8obsidian-wiki sessions-query "<topic>" --json
9```
10 
11If this exits 1 with "run sessions-build first", tell the user and offer `/session-brain`. If
12`graph.json` is more than ~7 days old, mention it and offer a rebuild — but **do not silently
13rebuild**, since that is a multi-second operation the user did not ask for.
14 
15## Step 2: Rank
16 
17The scoring already combines four signals, so take the ordering as given rather than re-ranking:
18 
19- **similarity** — TF-IDF cosine against the session's text
20- **cluster lift** — a session inside the best-matching topic scores higher even if its own words
21  never matched. This is why a session that never said "telemetry" can still surface for it.
22- **bookmark boost** — a human already flagged this session as worth keeping
23- **time decay** — 90-day half-life, applied with a floor so an old exact match still outranks a
24  fresh weak one
25 
26Useful filters: `--project NAME`, `--cluster N`, `--since DATE`, `--top N`.
27 
28## Step 3: Present
29 
30Show the top ~5 as a compact table — title, project, date, topic, and the `why` string, which
31already explains the match. Do not dump the raw JSON at the user.
32 
33Two things must be stated honestly rather than glossed over:
34 
35- Entries with `loadable: false` are **history-only**: the transcript has been pruned from disk
36…