Write weekly reports and standups with AI
Skills used in this guide (3)
跨 git worktree、分支或 PR 促成只读站会，比较变更并产出一份整合计划。
npx skills add thedotmack/claude-mem -g- npx skills add volcengine/OpenViking -g
数据驱动的 Scrum Master，运行 Python 脚本分析冲刺速度、健康度与回顾，给出预测与改进趋势。
npx skills add alirezarezvani/claude-skills -g
The pain of weekly reports isn't "can't write" — it's "turning a log into outcomes". The point: give AI the raw material (what you did, what's stuck), have it close into "progress / risk / next", not invent content for you.
Who this is for / not for
For:
- Weekly report / daily standup to sync
- Scattered notes (commits, chat) but lazy to polish
- Want a consistent structure, boss sees the point
Not for:
- Asking AI to invent things you didn't do (that's fabrication)
- Confidential progress (mask before external services)
- Fully zero-touch automated reporting
Before you start
- Material: this week's commits, task status, key chat excerpts.
- Template: company report format (progress / risk / next).
- Reader: boss / team, decides depth.
Steps
1. Install the skill
standup is a standup skill. Copy its folder into your client's skills directory (same as any Agent Skill):
| Client | Directory |
|---|---|
| Claude Code | ~/.claude/skills/ |
| Codex | ~/.codex/skills/ |
| Cursor | ~/.cursor/skills/ |
| Trae | ~/.trae/skills/ |
| CodeBuddy | ~/.codebuddy/skills/ |
| Tongyi Lingma | ~/.lingma/skills/ |
2. Feed material, fix structure
From this week's scattered notes, write a weekly report, fixed 3 parts:
Progress (by project, only what landed) / Risk & blockers (specific, with impact) / Next week (verifiable).
Compress log into outcomes; no "following up" — say where it landed and the result.3. Shorter standup version
Use standup for "yesterday / today / blocker" in three lines; blocker says who you need:
From my context, extract standup 3 lines: yesterday / today / blocked, who to help.
Each under two lines.4. You finalize
Review AI's draft, delete fabricated progress, keep only what really happened.
Pitfalls
| Pit | Symptom | Fix |
|---|---|---|
| Invented | Wrote things not done | Mark absent material [TBD] |
| Log style | No point | Force progress/risk/next |
| No result | "following up" filler | Require where it landed |
| Leak | Sensitive out | Mask confidential first |
| Too vague | Can't tell what | Each with verifiable evidence |
FAQ
No notes, can it write?
Hard. Best with commits / task status as material; pure memory → AI fabricates.
One template for standup + weekly?
Standup 3 lines shorter; weekly 3 parts fuller; switch depth by reader.
Team-wide consistency?
Give one template, everyone outputs same structure, then just concatenate.
AI exaggerates?
Yes — you delete fabricated parts on finalize, keep real only.
Scheduled generation?
Yes — make "collect material + generate" a scheduled task, draft for you to confirm.
Related guides
Continue across these scenes:
- Handle your inbox and support messages with AI
- Turn meeting recordings into a task list with AI
- Build dashboards and visualizations with AI
相关源码
- thedotmack/claude-mem · standup source
- volcengine/OpenViking · daily-report source
- alirezarezvani/claude-skills · scrum-master source
- Agent Skill open format spec
一句话结论
结论：周报要把流水账压成产出——喂原料、定"进展/风险/下一步"三段，AI 编出来的部分你定稿时删掉。
References
Related Scenes
Write PRDs and product specs with AI
The pain of PRD isn't "wrong format" — it's "goal and acceptance unclear, dev guesses". The point: give background, target user, problem; let AI output "feature list + acceptance + edges", not decide "why build" for you.
Translate and localize docs with AI
Machine translation's scare is "fluent but distorted" — wrong terms, wrong tone, stiff local phrasing. The point: give a glossary and target-reader tone, have it "translate + keep proper nouns", then you verify key sentences — not trust the whole blindly.
Handle your inbox and support messages with AI
The red line for email automation is "draft yes, auto-send careful". "Triage + draft + you hit send" is safest; letting AI blast-send means you own the mistake, not it.
Turn meeting recordings into a task list with AI
The biggest waste in meetings isn't the meeting — it's nobody remembering "who, by when, what" afterward. The point of this flow is to compress scattered discussion into a task table with owners and due dates, not to write another minutes doc nobody reads.