session-compare

1# Session Compare
2 
3Compare two Claude Code sessions side-by-side using Agent Monitor data.
4 
5## Input
6 
7The user provides: **$ARGUMENTS**
8 
9This may be:
10- Two session IDs: "abc123 def456"
11- "best vs worst" — compare highest and lowest productivity sessions
12- "latest 2" — compare the two most recent sessions
13- A session ID + "vs average" — compare one session against the baseline
14 
15## Procedure
16 
171. **Identify sessions to compare**:
18   - If two IDs given: fetch both from `http://localhost:4820/api/sessions/{id}`
19   - If "best vs worst": fetch sessions, score by completion + cost efficiency, pick extremes
20   - If "latest 2": `GET /api/sessions?limit=2` (default sort: most recently updated first)
21   - If "vs average": fetch session + compute averages from last 50 sessions
22 
232. **Gather detailed data** for each session:
24   - Session metadata: `GET /api/sessions/{id}`
25   - Events: `GET /api/events?session_id={id}`
26   - Agents: `GET /api/agents?session_id={id}`
27   - Cost: `GET /api/pricing/cost/{id}`
28 
293. **Build comparison**:
30 
31   ### Overview Comparison
32   | Metric | Session A | Session B | Difference |
33   |--------|-----------|-----------|-----------|
34   | Status | completed | error | — |
35   | Model | sonnet-4 | sonnet-4 | same |
36   | Duration | 12m 34s | 45m 12s | +32m 38s |
37   | Total Cost | $0.0234 | $0.1456 | +522% |
38   | Events | 45 | 187 | +315% |
39   | Tools Used | 8 | 12 | +4 |
40   | Error Count | 0 | 7 | +7 |
41   | Agents | 2 | 5 | +3 |
42 
43   ### Token Comparison
44…