daily-status

name	daily-status
description	Generate a daily status update for any squad's Slack workflow. Pulls recent git activity, open PRs, and blockers to produce a formatted update matching the team's standup template.

Daily Status Update Generator

Generate a status update for a squad Slack standup thread. Adapts to whatever format the team uses (P/B/R, Yesterday/Today/Blockers, custom).

Arguments

The user may pass arguments to customize the output:

• /daily-status — default P/B/R format, last 24 hours, current repo
• /daily-status since Friday — custom time window
• /daily-status --repos torch-spyre,spyre-inference — multi-repo
• /daily-status --format ytb — Yesterday/Today/Blockers format
• /daily-status --label vLLM — filter to issues with a specific label

Steps

1. Identify the user by running git config user.name and gh api user --jq .login.

2. The repo scope defaults to the current repo from gh repo view --json nameWithOwner --jq .nameWithOwner. If the user specifies --repos, iterate over each. If the user specifies an org, query all repos under that org.

3. Use P/B/R format unless the user specifies otherwise:

   Format flag	Sections
   pbr (default)	Progress / Blockers / Review
   ytb	Yesterday / Today / Blockers
   custom	Ask the user for section names

4. The time window defaults to "last 24 hours". Parse natural language like "since Friday", "this week", "last 3 days".

5. Gather activity by running the following, scoped to the time window and repo list:

   # Commits authored in the time window (across all branches)
   git log --all --author="$(git config user.name)" --since="<WINDOW>" \
     --format="%h %s" --no-merges
   
   # PRs authored (open or merged in the window)
   gh pr list --author="@me" --state all --limit 20 \
     --json number,title,state,updatedAt,headRefName \
     --jq '.[] | select(.updatedAt > "<ISO_TIMESTAMP>") | "\(.state) #\(.number): \(.title)"'
   
   # Issues assigned to the user (optionally filtered by label)
   gh issue list --assignee="@me" --state open --limit 20 \
     --json number,title,labels,updatedAt \
     --jq '.[] | "#\(.number): \(.title) [\(.labels | map(.name) | join(", "))]"'
   

   If --label is specified, add --label "<LABEL>" to the issue query.

6. Look for blockers:

   # PRs waiting on external review
   gh pr list --author="@me" --state open --limit 10 \
     --json number,title,reviewDecision,updatedAt \
     --jq '.[] | select(.reviewDecision != "APPROVED") | "#\(.number): \(.title) (review: \(.reviewDecision // "PENDING"))"'
   
   # CI failures on open PRs
   gh pr list --author="@me" --state open --limit 10 \
     --json number,title,statusCheckRollup \
     --jq '.[] | select(.statusCheckRollup | map(select(.conclusion == "FAILURE")) | length > 0) | "#\(.number): \(.title) — CI failing"'
   

   If nothing is blocking, write "None".

7. Find PRs where feedback was addressed and reviewers need to re-check:

   # PRs with changes requested that were pushed to recently
   gh pr list --author="@me" --state open --limit 10 \
     --json number,title,reviewDecision,updatedAt \
     --jq '.[] | select(.reviewDecision == "CHANGES_REQUESTED") | "#\(.number): \(.title)"'
   

8. Format the output as a message matching the chosen template, ready to paste into Slack:

   P/B/R format (default)

   **P:**
   - <progress bullets>
   
   **B:**
   - <blocker bullets or "None">
   
   **R:**
   - <review bullets or "None">
   

   Yesterday/Today/Blockers format

   **Yesterday:**
   - <what was done>
   
   **Today:**
   - <planned work — infer from open issues/PRs in progress>
   
   **Blockers:**
   - <blockers or "None">
   

   Rules:

   • One line per bullet. Use PR/issue numbers as references (e.g., "#1234").
   • No emojis unless the user requests them.
   • Keep it concise — the reader skims dozens of these in a thread.

9. Show the formatted update and ask if the user wants to adjust anything before posting.

Notes

• If gh is not authenticated, instruct the user to run gh auth login.
• For multi-repo updates, group bullets by repo only if there are items from more than one repo. Otherwise skip the repo header.
• The skill works in any GitHub-hosted project, not just torch-spyre. It reads the repo context from the current working directory.
• The output is scoped to whoever is running it via their git/gh identity. No squad configuration needed.