1on1-analysis

If vault_path is not in context, read ~/.myna/config.yaml first. If the file does not exist, tell the user to run /myna:setup and stop.

1:1 Pattern Analysis

Reviews 1:1 session notes for a specific person and surfaces factual patterns across sessions. Read-only — inline output only.

Strict data boundary: Report only what's in the notes. No inferences about relationship quality, engagement, or morale. "Fewer topics" is a count, not a signal about anything.

---

Resolve the Person

Match the user's input against people.yaml via fuzzy resolution. If multiple matches, ask the user to clarify. Proceed for any relationship tier — the user may analyze 1:1s with their manager or a peer.

---

What to Read

Primary source: Meetings/1-1s/{person-slug}.md

Sessions are typically organized with a dated heading, a prep section (checklist items), and a notes section (discussion, action items, decisions). Adapt to whatever structure is actually present — the user may have added custom sections or the format may have evolved.

Read all sessions. Parse all ## {YYYY-MM-DD} Session headings and sort ascending by date before analysis.

Default to sessions from the last 12 weeks. The user can specify a different range ("last 6 months", "last 10 sessions").

Additional sources for action item follow-through:

• Projects/*.md — Grep for open and completed tasks with [person:: [[{name}]] involving this person. Completed tasks (- [x]) provide additional evidence that action items were addressed, even if not reflected in meeting notes.

Note on two distinct item types:

• Prep items — checkbox items associated with the session. Checkbox state ([x] = addressed, [ ] = not addressed/carried forward) is the primary signal for carry-forward analysis.
• Action items — commitments made during the session, typically under a notes or action items section. These are freeform text. Track them by checking whether the same item text appears in a later session as addressed.

---

What to Analyze

1. Action Item Follow-Through

For each session (starting from the second), check whether action items from the previous session appear as addressed in the current session's notes or prep.

• Count action items from the previous session that were addressed vs. not addressed
• List unaddressed items specifically: the item text and the session it came from
• Calculate follow-through rate: X of Y action items addressed across analyzed sessions

2. Recurring Unresolved Topics

Scan discussion notes across all sessions for topics that appear in 3 or more sessions. A topic recurs when the same term, phrase, or subject appears in the discussion text across sessions.

A topic is resolved if: the corresponding prep item is checked [x] in the same or a subsequent session, or the same or a subsequent session contains "resolved", "closed", "done", "decided", or "won't pursue" in close proximity to that topic text.

List each recurring unresolved topic:

• Topic name/description
• Sessions it appeared in (dates)
• Whether it was ever marked resolved

Do not infer why it recurs — just report the facts.

3. Carry-Forward Rate

Carry-forward is tracked via Prep checkboxes. For each session, count:

• Prep items marked - [ ] (not addressed — carried forward or dropped)
• Prep items marked - [x] (addressed in that session)

If a - [ ] item from session N appears again in session N+1's Prep, it is confirmed as carried forward. If it disappears, it was dropped.

Carry-forward rate per session = unchecked items / total prep items. Report the per-session rate and the average across sessions.

---

Output Structure

## 1:1 Analysis — [Person Name]

**Sessions analyzed:** [N] sessions ([date range])  **File:** [[Meetings/1-1s/{person-slug}]]

---

### Summary
[3-4 sentences covering: overall follow-through rate, whether any recurring topics are a concern, and carry-forward trend. Use fewer sentences if the facts are sparse. Write a real synthesis — what the pattern actually means in practice, what stands out, what warrants attention — not a restatement of the counts below. Apply the same strict data boundary: synthesize patterns, not assessments. Do not infer relationship quality, engagement, or morale — report only what the data shows.]

### Action Item Follow-Through

**Overall rate:** [X]% ([Y of Z] action items addressed across analyzed sessions)

**All analyzed sessions:**
| Session | Action Items | Addressed | Not Addressed |
|---------|-------------|-----------|---------------|
| Apr 2   | 3           | 2         | 1             |
| ...     | ...         | ...       | ...           |

**Unaddressed items (from analyzed sessions):**
- "Review caching architecture proposal" — from Mar 21, not addressed in Mar 28 or Apr 2
- "Send growth plan template" — from Apr 2, not yet addressed

---

### Recurring Unresolved Topics

Topics appearing in 3+ sessions without resolution:

- **API spec ownership** — appeared in: Feb 14, Mar 7, Mar 21, Apr 2 (4 sessions). Not resolved.
- **Q3 growth goals discussion** — appeared in: Mar 7, Mar 28, Apr 2 (3 sessions). Not resolved.

---

### Carry-Forward Rate

**Average carry-forward rate:** [X]% (average across [N] sessions)

| Session | Total Prep Items | Unchecked | Checked | Carry-Forward % |
|---------|-----------------|-----------|---------|-----------------|
| Apr 2   | 6               | 3         | 3       | 50%             |
| ...     | ...             | ...       | ...     | ...             |

Edge Cases

• No 1:1 file found: "No 1:1 meeting file found for [person]. 1:1 analysis requires meeting notes in Meetings/1-1s/{person-slug}.md."
• Fewer than 3 sessions: Show what's available, note "Analysis is limited — only [N] sessions found."
• Notes section empty in a session: That session counts toward the total but cannot contribute to action item or recurring topic analysis. Note how many sessions had no notes.
• No action items in any session: Skip the Action Item Follow-Through section, note that no action items were found in the notes.
• No prep checkboxes: Skip Carry-Forward Rate, note that no prep items were found.
• User specifies a range: Filter sessions by date range or count, proceed with the filtered set.