plan-week

name	plan-week
description	Weekly planning assistant. Reviews all PARA areas by priority, pulls Linear backlog, checks active experiments, and produces a prioritized action plan. Use when user says "plan my week", "weekly plan", "what should I focus on", "Monday planning", "plan the week", "weekly priorities", or starts the week and wants to get organized. Also triggers on "mid-week refresh" or "end of week review".
allowed-tools	["Read","Write","Edit","Bash","Glob","Grep","WebFetch","Agent","mcp__Linear__list_issues","mcp__Linear__list_teams","mcp__Linear__list_projects","mcp__Linear__list_cycles","mcp__Linear__get_issue","mcp__Linear__get_project","mcp__MyNotes__extract_todos","mcp__MyNotes__search_ai_content"]

Plan Week

Generate a prioritized weekly plan by scanning all responsibility areas, pulling in Linear work, and ensuring high-priority areas have concrete actions scheduled.

When triggered, determine the mode from context:

User says	Mode
"plan my week", "weekly plan", "Monday planning"	Full plan (default)
"refresh the plan", "update weekly plan", "mid-week check"	refresh
"how'd the week go", "weekly review", "end of week"	review

Arguments

$ARGUMENTS

• No arguments: full weekly plan for the current week.
• refresh: re-run mid-week to update priorities based on what's changed.
• review: end-of-week review — what got done, what carried over.

Key Behaviors

1. Always start by reading the last weekly plan to establish continuity.
2. Gather context in parallel — don't sequentially read every area.
3. Be opinionated about priorities — the area registry has explicit H/M/L ratings. Don't let the user skip H items without acknowledging the skip.
4. CNW experiments and supply chain security are non-negotiable weekly items.
5. Keep the plan concise — 5-7 Must Do items max. This is a focusing tool, not a dump of everything possible.
6. Reference existing commands — suggest running /cnw-stats-analyzer, /metrics-review, /kudos check etc. rather than inlining their logic.

Area Registry

Each area has a priority level that determines how it's handled in weekly planning.

Priority levels:

• H (High): MUST have at least one concrete action this week. Flagged red if missing.
• M (Medium): Should have an action OR a conscious "nothing needed this week" decision.
• L (Low): Quick glance — only escalate if something is on fire.

High Priority

Area	Path	Weekly Action Required
Onboarding & Activation (Quark-a)	onboarding-activation/	Twice-weekly metrics pulse (Mon full, Thu refresh). Pull numbers against 550 new workspaces / $3.3M ARR targets and 600 claimed Cloud workspaces/wk tactical target. See "Onboarding & Activation pulse" section below.
CNW Stats & Experiments	cnw-stats/	Run /cnw-stats-analyzer, review experiment results, check targets (200 Cloud yes/day, 2K starts/day). Plan or adjust experiments. Feeds into Onboarding & Activation pulse.
Supply Chain Security	scan-audit/scans/supply-chain-security/	Ensure /scan-and-audit ran this week with supply-chain included. Review findings. Weekly cadence — don't let this slip.

Medium Priority

Area	Path	Weekly Check
Metrics Review	metrics-review/	Check npm trends, docs traffic direction. Run /metrics-review npm at minimum.
Project Health	project-health/	Scan for blocked/zombie projects in Linear. Any project overdue?
Competitor Intel	competitor-intel/	Glance at latest scan. Anything requiring a response this week?
SPACE / Productivity	space-metrics/	Engineering productivity concerns? Cycle times trending wrong? Blocked work?

Low Priority

Area	Path	When to Escalate
Personnel	personnel/	Escalate if: review cycle approaching, someone flagged concerns in 1:1
Syncs	syncs/	Prep happens naturally before each sync — no weekly action unless topics accumulating
Team Capacity	team-capacity/	Escalate if: someone overloaded, key person OOO, hiring gap
Promotions	promotions/	Escalate if: promo cycle deadline approaching
Customer Deps	customer-deps/	Escalate if: key customer escalation or churn risk
Community Sentiment	community-sentiment/	Escalate if: viral negative thread, major complaint pattern
Scan & Audit	scan-audit/	Runs weekly — just verify it ran. Supply chain findings bubble up to H. Sub-scans: AI landscape, frameworks, runtimes, API surface, dependency health.
TOIL	time-off-in-lieu/	Annual — December only

Planning Flow

Step 1: Context Gathering (parallel)

Run these in parallel to gather state:

1. Last week's plan: Read .ai/para/areas/weekly-plans/ for the most recent plan. Check what was planned vs what got done.

2. TODO.md: Read .ai/TODO.md for in-progress and pending items.

3. Linear current cycle: Query active cycles for Nx CLI and related teams. Pull issues assigned to Jack or his direct reports that are in-progress or planned for this week.

4. CNW experiment status: Read .ai/para/areas/cnw-stats/ for active experiments and latest stats. Are we hitting targets?

5. Onboarding & Activation pulse: Read .ai/para/areas/onboarding-activation/ for the latest pulse. Pull the numbers listed in the pulse section below — this is the primary revenue-linked signal and must not be skipped.

6. Scan-audit state: Read .ai/para/areas/scan-audit/state.json — when did scans last run? Any overdue?

7. Upcoming syncs: Check .ai/para/areas/syncs/ for topics accumulating across team syncs.

8. 1:1 action items: Scan .ai/para/areas/personnel/*.md for open action items from recent 1:1s.

9. Promotions/reviews: Check if any review cycle deadlines are within 4 weeks.

Step 1b: Onboarding & Activation pulse

Cadence: twice-weekly — Monday (full) in every new plan, Thursday (refresh) via /plan-week refresh. Once-weekly only acceptable when on PTO.

This is the revenue-linked pulse. Save to .ai/para/areas/onboarding-activation/pulse/{YYYY-MM-DD}.md and link from the weekly plan. If the prior pulse is less than 48 hours old, copy it forward and diff — don't re-run every source.

Numbers to capture:

Number	Source	Target
Claimed Cloud workspaces (last 7d)	Nx Cloud telemetry (ask Cory/Nicole if no direct query yet)	600/wk
Pace toward 550 new workspaces (period-to-date)	Company self-serve dashboard	550 cumulative
Pace toward $3.3M new ARR (period-to-date)	Revenue dashboard (ask Cory)	$3.3M cumulative
CNW completions (7d avg/day)	/cnw-stats-analyzer	baseline ~1,500/day
nx init completions (7d avg/day)	/cnw-stats-analyzer	baseline ~250/day, stretch 300/day
nx import completions (7d avg/day)	/cnw-stats-analyzer	trend only
Cloud opt-in rate from CNW	/cnw-stats-analyzer	20% (currently ~12%)
Claim rate (init vs CNW)	Cloud telemetry + CNW stats	pull weekly
Stage conversion rates	Cory's PLG dashboard	trend only — flag regressions
Active experiments status	areas/cnw-stats/ + Quark-a notes	decisions this week

What the pulse produces:

• One-line verdict: on track / at risk / off track for the 600/wk tactical target and the 550/$3.3M company targets.
• Deltas vs the prior pulse (which numbers moved, in what direction).
• Blockers or surprising signals worth raising in the Quark-a sync.
• Decisions needed this week (e.g. "launch variant B", "cut experiment X", "ask Nicole for X dashboard").

Source coverage note: CNW stats are queryable today. Cloud-side numbers (claimed workspaces, ARR pace) may not have a direct query path yet — in that case, capture what's accessible and flag "needs Cory/Nicole input" explicitly. Don't fabricate numbers to fill the table.

Step 2: Area-by-Area Assessment

For each area in the registry, produce a one-line status and proposed action:

### [AREA NAME] — [H/M/L] — [STATUS EMOJI]

**Status**: One-line summary of current state.
**Last touched**: When this area was last reviewed/actioned.
**This week**: Proposed action OR "No action needed — [reason]".

Status emojis:

• on-track — healthy, no concerns
• needs-attention — something to address this week
• at-risk — overdue or trending wrong, must act
• blocked — can't progress without external input

Step 3: Compile Weekly Plan

Group actions by day preference (not strict time-blocking, just suggested flow):

# Weekly Plan — {WEEK_START} to {WEEK_END}

## Carry-Over from Last Week
- [ ] Items that didn't get done (with reason)

## This Week's Focus

### Must Do (from H areas + escalated items)
- [ ] Onboarding & Activation: Monday pulse — pull all numbers in the pulse table, save to `onboarding-activation/pulse/{date}.md`, produce verdict
- [ ] Onboarding & Activation: Thursday refresh pulse (during `/plan-week refresh`)
- [ ] CNW: [specific action — e.g., "Review experiment X results, decide on next variant"]
- [ ] CNW: [specific action — e.g., "Check daily starts and Cloud opt-in against targets"]
- [ ] Supply Chain: [specific action — e.g., "Verify scan-audit ran, review supply-chain findings"]
- [ ] [Any escalated items from M/L areas]

### Should Do (from M areas)
- [ ] Metrics: [specific action]
- [ ] Project Health: [specific action]
- [ ] Competitor Intel: [specific action]
- [ ] SPACE: [specific action]

### If Time Permits (from L areas with pending items)
- [ ] [Only items that are actually pending, not routine]

## Syncs & 1:1s This Week
- [ ] [Day]: [Sync name] — Topics: [from upcoming sync sections]
- [ ] [Day]: 1:1 with [name] — Follow up: [from action items]

## Active Experiments
| Experiment | Area | Status | Action This Week |
|-----------|------|--------|-----------------|
| [name] | CNW | [running/paused/needs decision] | [what to do] |

## Decisions Needed
- [ ] [Any pending decisions surfaced from Linear, syncs, or area reviews]

## End-of-Week Checklist
- [ ] Onboarding & Activation: Monday + Thursday pulse both captured
- [ ] CNW targets checked (starts/day, Cloud opt-in/day)
- [ ] Supply chain scan verified
- [ ] Weekly plan items marked done or carried over
- [ ] Any kudos to capture? (`/kudos check`)

Step 4: Save & Confirm

1. Save the plan to .ai/para/areas/weekly-plans/{WEEK_START}.md
2. Create the directory if it doesn't exist.
3. Update TODO.md with any new action items from the plan.
4. Print a concise summary to the console.

End-of-Week Review (/plan-week review)

1. Read this week's plan.
2. For each item, check if it's done (from TODO.md, git log, Linear state).
3. Produce a scorecard:

## Week Review — {WEEK_START}

### Scorecard
- Must Do: 4/5 completed
- Should Do: 2/4 completed
- Experiments: 1 decision made, 1 still running

### Carried Over
- [ ] [Items moving to next week with reason]

### Wins
- [Notable accomplishments worth remembering]

### Misses
- [What didn't happen and why]

4. Save to .ai/para/areas/weekly-plans/{WEEK_START}-review.md
5. Feed carry-over items into next week's plan.

Mid-Week Refresh (/plan-week refresh)

1. Read current plan.
2. Run the Onboarding & Activation pulse again (Thursday touch). Save to onboarding-activation/pulse/{date}.md and diff vs Monday's pulse. If any metric shifted meaningfully, flag it in the updated plan.
3. Check what's done, what's blocked, what's new.
4. Re-prioritize remaining items.
5. Update the plan file in place.

Important Notes

• Onboarding & Activation pulse is non-negotiable — twice weekly (Mon full
  • Thu refresh). It's the revenue-linked signal for the 550 workspaces / $3.3M ARR company targets. If a number isn't queryable yet (Cloud-side), still run the pulse and flag the gap — don't skip the ritual.
• CNW experiments are non-negotiable — every week must have at least one experiment-related action (launch, review, decide, iterate).
• Supply chain is non-negotiable — weekly verification that scans ran and findings reviewed. This is the one scan-audit sub-report that must not slip.
• Don't overload the plan — 5-7 "Must Do" items max. If more, force-rank.
• "No action needed" is a valid answer for M/L areas — but it must be a conscious decision, not a skip.
• Pull from Linear, don't duplicate — reference Linear issue IDs, don't rewrite descriptions.
• The plan is a living doc — update it mid-week via refresh, don't let it go stale.