Index
Find and catalog past sessions across .stoobz/ directories.
Process
- •
Determine search scope:
- •Default: scan the nearest
.stoobz/directory (project-level first, then~/.stoobz/) - •If user specifies "all" or "everything": scan both project and global
.stoobz/ - •If user provides a search term: filter results by topic/ticket
- •Default: scan the nearest
- •
Scan for session directories — Walk
.stoobz/looking for directories containing session artifacts (TLDR.md,CONTEXT_FOR_NEXT_SESSION.md,RETRO.md,PROMPT_LAB.md,HANDOFF.md). - •
For each directory found:
- •Read the first 5 lines of
TLDR.md(if present) for the title and date - •Note which artifacts exist
- •Note the most recent modification date
- •Read the first 5 lines of
- •
Present the index:
markdown
## Session Index — {scope}
| Directory | Date | Summary | Artifacts |
| ----------------------------- | ---------- | ------------------------------- | --------- |
| `ENG-22456/uat-investigation` | 2026-02-06 | Statusline + session kit skills | T R P |
| `ENG-22456/qa-phase` | 2026-01-28 | QA environment profiling | T C |
| `memory-leaks` | 2026-01-26 | BEAM memory investigation | T C R |
**Legend:** T=TLDR C=Context R=Retro P=Prompt-Lab H=Handoff
- •If user is searching, highlight matching results and show context from the TLDR.
Rules
- •Read only headers — Don't load full file contents. First 5 lines of TLDR.md is enough for the index.
- •Sort by date — Most recent first.
- •Fast — This is a lookup tool. Don't analyze, just catalog.
- •Suggest pickup — If a result looks like it has unfinished work (has
CONTEXT_FOR_NEXT_SESSION.md), note: "Has resume context — run/pickupfrom that directory." - •Present to the user directly, don't write a file (this is a query, not an artifact).