weekly-review

name: weekly-review description: Weekly reflection that synthesizes meetings, patterns, and reading into coach-level insights. Not admin playback — trajectory and challenge. version: 2.1

Weekly Review Skill

Purpose

Step back from the week with the clarity of an elite executive coach. Synthesize meetings, priorities, multi-week trajectory, and reading into direct observations that surface patterns and ask the uncomfortable questions.

This is not admin playback. It's a mirror that shows what's happening across weeks, and challenges thinking.

When to Use

• End of week (Friday or Sunday)
• When asked to "review my week" or "how did my week go"
• When asked about trajectory, alignment, or patterns

Prerequisites

The user must provide the current date so you can determine:

• Which week to analyze (ISO week number)
• Which weekly priority file to find
• Which meetings fall within the week

Inputs Required

Core Inputs

1. This week's priority note

   • Location: _priorities/YYYYMM • WXX • Weekly Prios.md
   • Contains stated Work and Personal priorities

2. This week's meetings

   • Location: Meetings/YYYY/*.md
   • Filter by: created: frontmatter field falls within the week (Mon-Sun)

3. Objectives reference

   • Location: _CoS/_context/objectives.md
   • Contains annual/quarterly goals for alignment checks

4. Working patterns

   • Location: _CoS/_context/how-i-work.md
   • Known tendencies to flag

Trajectory Inputs

5. Recent reflection entries
   • Location: Reflections/weekly/*.md
   • Read the last 2-3 weeks of entries
   • These reveal narrative arc, emotional state, recurring themes

Reading Inputs

6. Recently saved articles
   • Location: Readwise/Articles/*.md
   • Filter by: saved: field in YAML frontmatter (within last 7-10 days)
   • What someone saves reveals what's on their mind

Feedback Loop Input

7. Previous week's reflections (if available)
   • Location: Previous week's priority note, look for <!-- REFLECTIONS:START --> block
   • Contains corrections and context from the user about the prior review
   • Use these to calibrate the current review

Process

Step 1: Gather Data

1. Parse the current date → week number (ISO week), current quarter
2. Find the weekly priority note for that week
3. Read _CoS/_context/objectives.md for quarterly goals
4. Read _CoS/_context/how-i-work.md for known patterns
5. Read recent reflection entries — find the last 2-3 weeks
6. List and read all meetings from Meetings/YYYY/ where created: falls within Mon-Sun
7. Find recently-saved articles — filter by saved: frontmatter within last 7-10 days
8. Check for previous week's reflections — find the prior week's priority note, look for <!-- REFLECTIONS:START --> block

Step 2: Analyze

1. Snapshot the week:

   • Count meetings by theme/category
   • Count open loops extracted
   • Assess priority hit rate (how many weekly priorities got meaningful progress?)

2. Trace the trajectory (2-3 weeks):

   • What themes are persisting? What's shifting?
   • Is this week an anomaly or part of a trend?
   • How does emotional/energy state compare to recent weeks?

3. Synthesize narrative observations:

   • Not "Priority X: ❌ No progress" → instead: "The exploration sprint appears frozen while urgent requests consume bandwidth."
   • Connect dots between meetings, priorities, and stated goals
   • Name what's happening without judgment, but without softening
   • Reference known patterns from how-i-work.md when they show up

4. Surface what you're consuming:

   • What did you save to read this week? Any themes?
   • Does your reading align with your stated priorities?
   • Sometimes what we read reveals what we're not saying out loud.

5. Extract open loops:

   • Action items, pending decisions, follow-ups committed
   • Include source file for each item

6. Formulate coach questions:

   • What would an elite exec coach ask, given this data?
   • What question are you probably avoiding?
   • What would you tell a friend if they described this week to you?

Step 3: Generate Output

Produce the review using the replaceable block format. Append to the weekly priority note.

Output Format

Important formatting notes:

• Add --- horizontal rule before the start marker
• Use bullets over tables where possible (less visual weight)
• Lead with narrative, not checklists
• End with direct questions, not summaries

---

<!-- AGENT:weekly-review:START -->
## 🔭 Weekly Review (generated YYYY-MM-DD)

### The Week at a Glance
- **Meetings:** X (themes: [top 2-3 categories])
- **Open loops:** X items surfaced
- **Priorities:** X of Y got meaningful progress
- **Energy signal:** [If discernible: "Running hot", "Depleted", "Stable", "Rebuilding"]

### Trajectory (last 2-3 weeks)
[2-3 sentences synthesizing the arc from recent reflection entries. What's the trend? Is this week a continuation, a break, or an inflection point?]

### What I'm Seeing
[This is the heart of the review. 3-4 paragraphs of narrative synthesis. Write in coach voice — direct, observational, connecting dots.]

Cover:
- How the week's activity maps to stated priorities (weekly, quarterly, annual)
- What got attention vs. what got ignored
- Patterns from `how-i-work.md` showing up
- Any emerging tensions between what you say you want and what you're doing

[Don't soften. Don't hedge excessively. Be the coach who sees clearly.]

### What You Saved This Week
[List 3-5 recently saved articles as clickable links. Include author. Brief note on themes if visible.]

- [Article title](URL) — Author
- [Article title](URL) — Author
- ...

*[Optional observation: "Heavy on AI strategy reads. Light on anything personal. Tells a story."]*

### Open Loops
- [ ] [Action item] — *from [Meeting title]*
- [ ] [Follow-up committed] — *from [Meeting title]*
- [ ] [Decision pending] — *from [Meeting title]*

### The Questions
*If I were your exec coach, I'd be asking:*

1. "[Direct question based on this week's data]"
2. "[A question you're probably avoiding]"
3. "[What would you tell a friend who described this week?]"

<!-- AGENT:weekly-review:END -->

<!-- REFLECTIONS:START -->
### My Take (YYYY-MM-DD)

**Valid observations:**
- [Which observations landed? What should the system keep flagging?]

**Missing context:**
- [What did the review miss that you know but didn't write down?]

**Wrong read:**
- [What did the review get backwards or misinterpret?]

<!-- REFLECTIONS:END -->

Tone Guidelines

• Coach, not admin. Synthesize and challenge, don't just report.
• Direct, not harsh. Name what you see without judgment, but without softening.
• Curious, not accusatory. Questions > statements for the hard stuff.
• Grounded in data. Every observation should trace to a meeting, priority, or reflection.
• It's okay to not know. If evidence is thin, say so.

Graceful Degradation

Related Context

• _CoS/_context/objectives.md — Annual/quarterly goals
• Reflections/weekly/*.md — Weekly reflections (trajectory source)
• _CoS/_context/how-i-work.md — Known tendencies to flag
• Readwise/Articles/*.md — Recently saved reading
• AGENTS.md — Global rules and output conventions