| name | unfinished-threads |
| description | Conversations still in exploring or crystallizing stage with open questions at or above min_importance. Use when user feels there's unfinished work but doesn't know where. |
| layer | L2 |
| reads | ["L2.summaries"] |
| writes | [] |
| citations | required |
| determinism | pure-function |
| allowed-tools | mcp__my-brain__unfinished_threads |
unfinished-threads — L2 open-work surface
Framework context
L2 skill that filters to summaries matching:
thinking_stage IN ('exploring', 'crystallizing')importance IN ('breakthrough', 'significant') (or as configured)open_questions non-null and non-empty and not "none identified"
Orders by importance DESC, then msg_count DESC. Max 25 rows.
Distinct from L3 open-threads because this is single-summary granularity — one row per conversation — not domain-aggregated.
When to invoke
- User says "what's still unfinished" or "what's open that matters"
- Weekly review ritual
- Before planning sessions — surface what you owe yourself
Input
domain: str | None
importance: "breakthrough" | "significant" = "significant"
Output contract
## 🧵 Unfinished Threads {domain or 'across all domains'}
**Importance filter**: {importance}+ · **Results**: N
💎 **{title}** · {domain_primary} · {stage}
[conv_id · date] · {msg_count} msgs
> {summary 200ch}
Open questions (max 3 shown):
- {q 200ch}
- {q 200ch}
Does NOT do
- Aggregate across conversations (that's L3
open-threads)
- Close threads automatically
- Mark threads as abandoned (that's L3
dormant-contexts)
Verification checklist
