Standup Prep

Use this when you have ~2 minutes before standup and need a clear read on iteration status before walking into the room. Output is a personal reference sheet — not a script, not a read-aloud artifact.

---

Inputs

Paste any or all of the following. Raw is fine — no required format.

1. Iteration plan — stories in flight, their status, acceptance criteria if handy
2. Yesterday's updates — what shipped, what got stuck, what moved
3. New signals — Slack threads, PR comments, stakeholder messages since yesterday

Minimum: provide #1. The agent flags explicitly what it cannot assess without #2 or #3.

---

Steps

Step 1 — Parse

Auto-load yesterday's day-close handoff: Before parsing pasted inputs, check for journals/day-close-{{YESTERDAY}}.md. If it exists, load it as input #2 (Yesterday's updates) automatically. Kate does not need to paste it. If the file exists and Kate also pastes yesterday's updates, merge them -- the file may have items Kate's paste does not, and vice versa. If the file does not exist, proceed with whatever Kate pastes. Do not error.

Read all inputs and identify:

• Stories in flight and their last known status
• Explicit blockers or risks mentioned
• Signals of drift: story taking longer than expected, AC not started, PR sitting unreviewed, no update in 48+ hours
• Anything new the team may not have seen yet

Do not invent status. If a story has no update signal, mark it Unknown — no signal and move it to Questions to Ask.

Step 2 — Generate

Produce the following sections:

Story Status Table

Story	Status	Why
[title]	On Track / At Risk / Blocked / Done	[one sentence — specific signal, not vague]

Status definitions:

• On Track — work is progressing, no signals of delay, within expected pace
• At Risk — a specific signal suggests delay or drift (name the signal)
• Blocked — work cannot continue without external input (name what and who)
• Done — accepted and off the board
• Unknown — no signal available; raise in standup

Blockers to Raise

Format: [Story] is blocked on [what] — needs [who] to [action].

If none: write "No blockers." Do not manufacture them.

Questions to Ask

Specific questions the PM should raise in standup — about unknown statuses, stalled PRs, stories with no recent signal, or risks needing team input.

Format: [Specific question] — listening for [what].

Vague questions ("How's it going?") are not acceptable here.

Pace Check

One sentence: is the iteration on track to hit its goal at current velocity? If the data doesn't support a clear answer, say so explicitly.

Things You Might Not Have Seen (omit if empty)

Surface anything from new signals -- Slack, PRs, stakeholder messages -- that the team may not know yet. If nothing new was provided or found, omit this section entirely.

Housekeeping Targets (omit if the day is full of meetings or active project work)

If the day looks light (few meetings, no urgent commitments), suggest a housekeeping block with specific targets. Pull from:

• Stale continuation prompts (glob prompts/continuation-*.md, flag any >14 days old)
• Stalled inbox items (check inbox/_queue.md for items in raw/triaged >14 days or processing >7 days)
• Superseded continuation prompts that need deletion

Format: "Light day -- good candidate for a housekeeping block. Targets:" followed by a bulleted list of 2-4 specific items. If none of these sources have actionable items, omit this section entirely.

Step 3 — Flag Uncertainty

Before delivering output, review:

• Any status assigned without a clear signal? Reclassify as Unknown.
• Any blocker that looks like it may already be resolved? Flag as Monitor, not raise.
• Anything missing that the PM should check before walking in?

List these at the bottom under: Check before you go in.

---

Output Format

Personal reference sheet. Plain text or markdown. Designed to be glanced at, not read aloud. Keep it to one screen.

---

Writing Rules

• Don't invent. No signal = Unknown, not a best guess.
• Status must earn its label. Every non-On Track status needs a named, specific reason.
• Questions must be specific. Name the story, name what you're watching for.
• Flag uncertainty explicitly. Hedge language beats false confidence every time.
• This is not a standup script. Never produce read-aloud text or "you should say..." framing.

---

Frequency

Daily. Gets faster with practice as the PM learns what signals to paste. If Kate ran /daily-wrapup the evening before, yesterday's context is pre-loaded from journals/day-close-{{YESTERDAY}}.md. No paste needed for input #2.