Standup Digest
Turns yesterday's git activity into a concise standup update — what shipped, what's in progress, and any blockers — grouped and readable, not a raw commit dump.
What this skill does
Key features
- Summarises recent commits into themes, not a raw log
- Separates done / in-progress / blockers
- Keeps it short enough to read aloud in 30 seconds
- Flags nothing to report honestly instead of padding
Use cases
- Async daily standup updates for a team
- End-of-day summary for your own records
- Weekly digest by widening the time window
SKILL.md
The skill definition — its metadata and the exact instructions an agent follows. Copy it into a SKILL.md file in your agent's skills folder.
| name | standup-digest |
|---|---|
| description | Summarise recent git activity into a concise standup update. |
| allowed-tools | Bash(git log:*), Read |
| window | default: since yesterday |
Musts
- Read recent history with
git logover the requested window (default: since yesterday). - Group commits into themes/features, not a raw one-line-per-commit dump.
- Structure the output as: Done, In progress, Blockers.
- If there’s little or nothing to report, say so honestly rather than inflating it.
Guidelines
- Keep it to a few bullets per section — readable aloud in about 30 seconds.
- Write for a teammate, not a machine: themes and outcomes, not commit hashes.
- Only list a blocker if it’s real and actionable; don’t manufacture one.
Example
Done
- Shipped login rate-limiting (feat/auth)
- Fixed the CSV export encoding bug
In progress
- Refactoring the billing webhook handler
Blockers
- Waiting on staging Stripe keys to finish webhook testing
FAQ
Does it post to Slack automatically?
No — it drafts the text. Posting is a separate action you take, so you can edit first. Keeping a human between draft and broadcast avoids embarrassing auto-posts.
What if I committed nothing yesterday?
It says so plainly ('No commits since yesterday; work in progress on X') instead of padding the update with filler.
Before running any skill, read its instructions and the tools it's allowed to use, and test it on a safe target first. See LLM security best practices.