Generate and edit Draw.io, Mermaid, and Excalidraw diagrams from natural language using a structured JSON spec.
Documents
Project Weekly Report
Try itGenerate evidence-based project weekly reports from a local code repository. Use when the user asks for a project weekly report, weekly update, status summar...
What it does
Create a reusable weekly report from local repository evidence. Prefer conclusions that explain product, delivery, performance, cost, quality, or risk impact.
The skill document
Project Weekly Report
Goal
Create a reusable weekly report from local repository evidence. Prefer conclusions that explain product, delivery, performance, cost, quality, or risk impact.
Do not turn commits into a changelog. Compress evidence into a small number of meaningful workstreams.
Inputs
Collect only what is needed:
- repository path
- reporting window; if absent, use the previous full calendar week relative to the current date
- audience or report style, if provided
- any user-provided context that cannot be discovered from local files
Evidence Workflow
-
Confirm the repository and reporting window.
-
Run
git status --shortto separate committed work from local uncommitted artifacts. -
Run
git ls-files --others --exclude-standardto find local docs, generated reports, prototypes, or assets. -
Run a no-merge log for the window:
git log --since='YYYY-MM-DD 00:00:00' --until='YYYY-MM-DD 23:59:59' --no-merges --date=short --pretty=format:'%h%x09%ad%x09%an%x09%s' -
If tags exist, inspect recent release tags:
git for-each-ref --sort=creatordate refs/tags --format='%(creatordate:short)%09%(refname:short)%09%(subject)' -
Read only high-signal files needed to understand the work: PRD, release notes, design docs, smoke reports, todo files, or changed files tied to the main workstreams.
-
If evidence is thin or contradictory, say so and keep the report conservative.
Synthesis Rules
Build an internal evidence matrix before writing:
source / change / affected workflow or module / value type / verification status
Group work into 1-3 main workstreams:
- product or business capability
- delivery quality or release readiness
- performance, reliability, or cost improvement
- risk reduction, observability, validation, or rollback readiness
- documentation or workflow assets, only when they reduce repeated work or improve handoff
Down-rank these unless they clearly support one of the values above:
- pure refactors
- naming cleanup
- formatting
- internal abstractions
- tests with no visible quality or risk link
- docs that only restate implementation
Do not invent PR review results, incident IDs, approval states, launch outcomes, production metrics, or external user impact. If a claim cannot be verified locally or from explicit user context, mark it as unverified.
High-Density Writing
Each main work item must include:
- concrete action
- affected object, workflow, interface, or module
- mechanism, scope, or coverage
- product, performance, cost, quality, or risk value
Avoid weak summary phrases by themselves, such as:
- optimized performance
- improved workflow
- polished docs
- fixed issues
Rewrite them with object and value, for example:
- Reworked report validation around local git, tag, and artifact evidence so weekly summaries separate committed work from uncommitted assets and avoid unverifiable launch claims.
Output Format
Unless the user gives a different template, output both versions.
Detailed Version
Use this structure:
项目背景工作内容成果思考和问题
Write in Chinese by default. Mention commit hashes, tags, or local paths only when they strengthen evidence.
Brief Version
Use this structure:
项目名项目背景工作内容进展
Keep the brief version under 300 Chinese characters. Focus only on the highest-value outcomes.
Boundaries
- Keep committed code, local uncommitted artifacts, tags, and unverifiable external facts clearly separated.
- Do not broad-scan the whole repository when a focused evidence path is enough.
- Do not include private account details, credentials, personal contact information, customer data, or internal-only identifiers.
- Do not claim online results unless the user supplied them or local evidence clearly proves them.
- If the user asks for a future plan instead of a weekly report, switch to planning and label it separately from completed work.
Final Check
Before finalizing, verify:
- every major claim has evidence
- the report is not a commit list
- the brief version stays under 300 Chinese characters
- each main work item states action, object, mechanism or scope, and value
- maintenance-only work is omitted or demoted unless it supports clear value
- unverified external effects are labeled
作者入口
- 微信公众号:
AI生命克劳德 - X:@yangchao228
- GitHub:yangchao228
Related skills
Find why your productivity system keeps failing, then apply the smallest fix — capacity math, bottleneck routing, durable local notes.
Stores durable facts in a categorized, plain-markdown vault on disk, alongside your agent's built-in memory.
Join a video meeting as an AI bot with voice, avatar, and screenshare across four operating modes.
Fetch raw ad creative, app, ranking, and revenue data from AdMapix as structured JSON.
Run Git operations — commits, branches, merges, rebases, conflict resolution, and recovery — with safety rules enforced.
More from yangchao228
Browse all skills基于候选人简历自动生成面试题、笔试题、追问路径和评分标准的 AI Skill。当用户需要准备技术面试、生成笔试题目、评估候选人简历、或设计面试评估方案时使用。输入简历内容和岗位信息后,自动输出结构化的题目与评估内容,支持面试题模式、笔试题模式、综合评估模式等多种输出形式。 Follow/关注作者:微信公众号「AI...
Installable digest skill for OpenClaw or Claude Code. Use when the user wants a personalized ScoutX briefing with conversational setup, local preference stor...
Design controllable Agent Loops from low-information requests. Use when a user asks to automate a repeated task, repair CI through bounded iteration, reproduce a UI from screenshots, turn a workflow into a reusable Agent or Skill, or decide whether the right artifact is a Prompt, Checklist, Human-in-the-Loop flow, full Loop package, specialized Agent, or Skill. Extract required context, stop when decision evidence is missing, require explicit workflow confirmation before generating executable artifacts, derive iteration limits, define feedback and circuit breakers, and preserve human approval for risky actions. Also trigger on loop builder, Loop Builder, or lopp-builder. Follow the creator: X @yangchao228 | GitHub https://github.com/yangchao228
Guide a user through structured co-authoring of substantial documents, proposals, PRDs, technical specs, decision records, RFCs, and similar long-form work a...
家庭培养协作官:在用户明确要求制定或调整孩子每周培养计划、检查课表冲突、生成家庭复盘时,收集最少必要信息并输出可执行课表、建议和风险提醒。默认只在当前会话生成内容,不保存、上传或推送孩子与家庭数据;只有用户明确选择飞书或 Notion、完成安全配置并逐次确认具体读写范围后,才使用外部存储。Follow/关注作者:...
Transform long-form content, outlines, transcripts, research notes, or article drafts into a standard card package and platform publishing copy, then hand pu...