Weekly monday review

Weekly Monday Review

The Monday 30-minute "where are we" loop is the keystone of the parallel-projects discipline (workflows/parallel-projects.md Habit 2). Without it, all 7 projects compete for attention every day; with it, 2 win the week and the others get a 30-minute Friday touchpoint.

This skill scans the operator's project notes, applies the needle-mover ranking, and produces the four-section weekly classification. It is a focus filter, not a status report — the goal is to be honest about which projects are explicitly not getting attention this week so they stop owing the operator mental load.

When to Use

• It is Monday morning and the operator says "let's do the weekly review", "Monday review", or "what should I focus on this week".
• The operator runs the skill on a non-Monday with --force (the discipline is Monday but the mechanism is the same; e.g. coming back from a holiday on a Wednesday, the review still needs to happen once before focused work resumes).
• The operator just finished a launch and asks for a re-rank — major events shift the needle-mover calculus and the previous Monday's classification is stale.
• Auto-fires from the start of the operator's workspace if no review has been recorded in the last 8 days.
• ECC has dashboard-builder (builds dashboards as artefacts) and project-flow-ops (generic project ops). Neither produces the 2-of-N focus pick + paused-project status update specific to Habit 2; use this one for the Monday loop.

Inputs

• The path to the operator's project notes directory. Default ~/Brain/Projects/. Honour any override the user gives.
• The weekly capacity, default 2 focus + 3 Friday touchpoint + remaining paused per workflows/parallel-projects.md Habit 2. Honour any override (a holiday-shortened week may warrant 1 focus + 2 touchpoints).
• The needle-mover criterion. Default: revenue evidence pending or imminent milestone within the next 14 days ranks high; stale silent work (more than 4 days no commit, per Habit 3 third-rail rule) ranks lower because it needs a small visible step before more focus is warranted.
• An optional reading of recent activity for each project name from the prior week — ~/Brain/Daily/<date>.md notes that mention the project, or a /graphify query "<project> last week" against the Brain corpus. Recent activity is a tiebreaker when two projects look equally needle-moving from the notes alone.
• An optional --write-status flag from the user. Defaults off; when on, the skill offers to update the status: field in each note's frontmatter to reflect the week's classification (still requires explicit confirmation per project).

How It Works

1. Glob the project notes directory for *.md files. For each, parse the frontmatter and extract status, repo, url, stack, created. Skip any file whose status is archived or done. Carry forward paused and active files as the week's universe of candidates.
2. For each candidate, gather signals: open-tasks count from the unchecked checklist items, days-since-last-commit from git log -1 --format=%cd against the repo URL in the frontmatter, open-questions count, last decision timestamp from the decisions log, and any mentions of the project name in ~/Brain/Daily/*.md notes from the last 7 days (or a /graphify query against the Brain corpus if the operator has built it).
3. Apply the needle-mover ranking. Boost a project for: imminent revenue evidence (live URL with traffic, customer commitment in the next 14 days), recent mentions in daily notes, recent commits with feat: or ship: prefixes. Penalise a project for: silent for more than 4 days with no commit (Habit 3 third-rail), open questions outnumbering open tasks (means the project is stuck rather than executing), no live URL after more than 60 days (likely a research project that should be reframed).
4. Produce the four-section classification. ## Focus this week gets the top 2 by rank. ## Friday touchpoint gets the next 3. ## Paused gets every remaining active/paused candidate. ## Force-visible gets every project silent for more than 4 days regardless of rank — Habit 3 says the third-rail fires on time, not on rank.
5. For each project line, render - **<name>** — <one-line rationale> · <next concrete action>. The rationale names the signal that drove the ranking; the next action is the smallest visible step (commit, PR, lead-form change, customer reply) that moves the project this week.
6. Detect drift signals during the scan — repo URLs that 404, frontmatter status that contradicts recent git activity, decision-log entries older than 30 days on an active project. List these in a ## Drift detected block at the end and chain to obsidian-sync-helper for any project where drift looks load-bearing.
7. Offer one optional follow-up: write the new status: value into each project's frontmatter (status/focus, status/touchpoint, status/paused, status/force-visible). Read-only by default; require per-project confirmation before writing. Never batch-write all projects in one yes.

Output shape

## Focus this week (2)

- **Project Alpha** — live URL waiting on a paid-ad test; data lands Wednesday · ship the conversion-fix PR Monday morning
- **Project Bravo** — customer commit Friday; PR open since Saturday · land + deploy the PR before EOD Monday

## Friday touchpoint (3)

- **Project Charlie** — research-only this week, no shipping pressure · 30-min Friday content draft
- **Project Delta** — paused on infra hardening, no urgency · 30-min Friday Sentry review
- **Project Echo** — partner conversation pending · 30-min Friday inbox check

## Paused

- Project Foxtrot, Project Golf — status updated to paused, no work this week.

## Force-visible

- **Project Hotel** — silent 6 days, third-rail fires · push an interim README update before Wednesday EOD.

## Drift detected

- Project Foxtrot — frontmatter says active but no commits in 18 days. Suggest running `obsidian-sync-helper` on Foxtrot.

Anti-patterns

• Do not auto-write status fields without per-project confirmation. Habit 2 is operator-owned; the skill proposes, the operator confirms.
• Do not pick more than 2 focus projects regardless of how the rank shakes out. The cap is structural — relaxing it to 3 has produced a worse week every time the operator has tried.
• Do not skip the force-visible block when no project triggers it. An empty ## Force-visible section is the signal that the third-rail is quiet this week, which is itself information.
• Do not generate a status report when the operator asked for a focus filter. Long prose about each project's history is documentation drift; one rationale line + one action line is the bar.
• Do not pull in projects with status: done or status: archived to "round out" the list. Done projects are done.

Related

• workflows/parallel-projects.md — Habit 2 is the source-of-truth for the Monday loop and the 2-of-N cap; Habit 3 third-rail rules drive the force-visible logic.
• stack/obsidian-brain.md — the project-note frontmatter convention this skill reads from.
• skills/obsidian-sync-helper — chained when drift is detected during the scan.
• skills/solo-billing-monitor — adjacent weekly ritual; many operators run the cost rollup the same Monday.
• ECC dashboard-builder and project-flow-ops — disambiguated above; neither produces the focus filter.