Week Review

<!-- Generated from `.claude/skills/week-review/SKILL.md` by `scripts/generate-agents-skills.py`. Do not edit. -->

Execution mode

Run inline in the current conversation by default, so this work can see what the user has already discussed, decided, or settled this session. Do not fork merely because this skill was selected. Only run in the background when the user explicitly asks for a background run or the host has already obtained a specific background-work approval for this run.

Delegated gathering (large-vault scaling)

This skill stays inline as described above: it keeps session awareness, it asks the user the questions, and it owns every interactive step. What it does NOT do inline is the bulk read-gathering, which on a mature vault (hundreds of notes, thousands of indexed messages, a live calendar and multiple integrations) can be large enough to exhaust the main conversation before the useful work starts.

So the gathering phase is delegated to one general-purpose subagent via the Agent tool, using the self-contained prompt in this skill's AGENT_INSTRUCTIONS.md:

1. Read .agents/skills/week-review/AGENT_INSTRUCTIONS.md.
2. Substitute its placeholders ({{TARGET_DATE}}, {{WEEK_START_DATE}}, {{TARGET_DATE_PLUS_1}}, {{DAY_NAME}}, {{MONTH}}, {{DD}}, {{YYYY}}).
3. Call the Agent tool with subagent_type: "general-purpose", that prompt, and a short description.
4. Verify it wrote the synthesis to 00-Inbox/Weekly_Synthesis_YYYY-MM-DD.md, then run the interactive review from its structured findings.

The subagent inherits MCP connections, runs in its own context, and that context is freed when it completes, so only its findings reach this conversation.

Use AGENT_INSTRUCTIONS.md verbatim. Read the file and pass its content as the subagent prompt, substituting only the placeholders. Do NOT hand-write a replacement brief from what you already know about the week: that is how steps get silently dropped, and the omission looks complete because nothing errors. If context from this conversation is worth adding, APPEND it to the file's content; never substitute for it.

Two caveats that are load-bearing:

• Do not count on hooks for the subagent's writes. The hooks declared in this skill's own frontmatter belong to this skill's run, not the subagent's, and whether the repository-wide hooks in the host's lifecycle settings reach a subagent's tool calls is not something a skill should assume either way. Nothing in this skill's gathering depends on a hook; the subagent's writes must stand on their own.
• Always fall back. If the subagent fails, times out, or returns nothing usable, say so plainly and run the gathering inline from the same AGENT_INSTRUCTIONS.md. A missing subagent must never mean a missing result.

Stays inline: the whole interactive review, priority-by-priority assessment, pattern discussion, goal updates, career evidence capture, next week's priority confirmation, the Dex Inbox check (Step 0.5), clearing completed tasks from 03-Tasks/Tasks.md after confirm (the delegated summary names this step), and the /identity-snapshot run that follows the synthesis. The subagent gathers evidence; it does not make judgements.

Purpose

Create a synthesis of the week reviewing activity, progress, and what was accomplished. Uses concrete metrics, not vague percentages.

Read working_week.days in System/user-profile.yaml and treat its last working day as the natural review point.

Step 0: Process Unprocessed Meetings

Before gathering data, ensure all meetings from this week are in the vault by running /process-meetings. This pulls any unprocessed meetings from the meeting source (Otter.ai, Granola, etc.), creates meeting notes, updates person/company pages, and extracts tasks — so the weekly synthesis has complete data.

• If no new meetings are found, continue silently
• If meetings are processed, note the count for the synthesis
• Do NOT ask for a skill rating after this sub-step — save that for the end of the full review

Step 0.5: Dex Inbox Check (Phone Captures)

After processing meetings, check for tasks captured from phone that haven't been triaged:

Use: reminders_list_items(list_name="Dex Inbox")

If the tool is unavailable or errors (Apple Reminders phone-capture is optional and may not be set up on this machine): skip this step silently — do not surface an error for a feature the user never enabled.

If items found:

• Surface them: "📱 Phone captures not yet triaged (X items in Dex Inbox)"
• Run triage flow: infer pillar, confirm with user, create task, mark Reminder complete
• If user wants to defer: leave in Dex Inbox

If empty: Skip silently.

Data Sources

1. Task Progress

• 03-Tasks/Tasks.md — Task completion status
• 02-Week_Priorities/Week_Priorities.md — Weekly priorities

2. Project Activity

• 04-Projects/**/*.md — Modified this week

3. Meetings & People

• 00-Inbox/Meetings/*.md — Meeting notes from this week
• People/**/*.md — Person pages updated

4. Learnings

• 06-Resources/Learnings/**/*.md — Explicit learnings
• System/Session_Learnings/*.md — Auto-captured session learnings

5. Daily Plans & Reviews

• Resolved daily-plan folder (System/folder-paths.yaml daily_plans, else the folder that already holds dated daily plans, else 00-Inbox/Daily_Plans/YYYY-MM-DD.md) — This week's daily plans (primary record of planning ritual)
• 07-Archives/Reviews/Daily_Review_YYYY-MM-DD.md — This week's reviews

6. Journals (If Enabled)

• 00-Inbox/Journals/YYYY/MM-Month/ — Morning/evening journals

Analysis Process

1. Weekly Priority Completion (Concrete, Not Percentages)

Don't say: "Goal X went from 40% to 55%" Do say: "You completed 2 of 3 weekly priorities"

Use: get_week_progress()

For each weekly priority:

• Complete: ✅ What was the deliverable? When did you finish?
• In Progress: 🔄 What specifically got done? What's left?
• Not Started: ❌ Why? Should it carry forward?

Surface concrete accomplishments:

"This week's priorities:

1. ✅ Ship pricing page — Complete (pushed to prod Wednesday)

   • Deliverable: New pricing page live at /pricing
   • Tasks completed: 5 of 5

2. 🔄 Write Q1 strategy doc — 60% complete

   • Done: Outline, competitive analysis, recommendations
   • Remaining: Executive summary, financial projections
   • 2 tasks left

3. ❌ Customer interviews — Not started

   • Reason: Calendar was too stacked
   • Recommendation: Carry to next week with protected time"

1.5 Semantic Goal-to-Work Mapping (if QMD available)

Check if semantic search is available by looking for qmd in PATH.

If available, enhance the weekly priority review with meaning-based analysis:

1. Auto-detect goal contributions: For each completed task this week, search:

   qmd query "task title/description" --limit 3

   against quarterly goals. Catch tasks that advanced goals without explicit links.

   • Example: "Built customer health dashboard" semantically matches goal "Improve NPS tracking" — different words, same work.

2. Cross-priority connections: Search for work that bridges multiple priorities:

   qmd query "priority 1 description" --limit 5

   Surface tasks that contributed to more than one priority.

3. Thematic patterns: Search for recurring themes across the week's work:

   qmd query "common theme from meetings/tasks" --limit 5

   Detect patterns like "most of your work this week clustered around customer retention" even when tasks used different terminology.

Integration: Merge findings into the Quarterly Goals table — add a "Hidden contributions" row for semantically-detected but not explicitly-linked work. Only show genuinely new connections, not things already captured by keyword matching.

If QMD unavailable: Skip silently. Task completion stats still work fine.

2. Task Completion Stats (Concrete Numbers)

Scan 03-Tasks/Tasks.md for completion timestamps from this week:

• Count tasks completed (look for ✅ YYYY-MM-DD in date range)
• Count tasks added mid-week
• Count tasks carried over

Surface:

"Tasks this week:

• Completed: 14 tasks
• Added mid-week: 6 tasks (scope creep?)
• Carried over: 3 tasks

Completion rate: 82% (14 of 17 planned)"

3. Quarterly Goals Progress (Concrete Milestones)

Don't use fake percentages. Use milestone counts and specific accomplishments.

Use: get_quarterly_goals()
Use: get_goal_status(goal_id) for each goal that has a goal_id

Skip get_goal_status for any goal whose goal_id is null, or that is marked provisional — there is no ID to look up, and the call will say so rather than return a status.

Check activity_known before judging any goal. When it is false, linked_priorities and linked_priorities_count are null, meaning not known, never none. Never report such a goal as stalled or as having no linked work; say its links cannot be read yet and pass on the recommendation about giving it an ID.

For each goal:

• Milestones completed this week
• Total milestones done vs. total
• Weeks since last milestone
• Specific accomplishments that moved the goal

"Quarterly Goals Progress:

Goal	Milestones	This Week	Status
Launch v2.0	3 of 5	+1 (Pricing page shipped)	On track
Improve NPS	1 of 4	No change	⚠️ Stalled (3 weeks)
Team Capacity	2 of 3	No change	On track

Goal 1 advanced because you completed Priority 1. Goal 2 needs attention — no linked work completed this week."