Preface

Submitting weekly reports on Friday afternoons often involves more than one repository: the main project, script libraries, and documentation repos each have their own commits. Manually scrolling through git log and categorizing them into “Completed This Week / In Progress / Risks” is both slow and prone to omissions. Common workarounds include using fixed templates plus manual copy-paste, or writing one-off scripts that only cover a single directory.

Below is an introduction to weekly-report on SkillHub (coordinate @user_1f530050/weekly-report-new): after being triggered via natural language in an Agent, it scans Git commit records across projects in a specified directory, organizes them into a formatted weekly report, and writes it to a specified path. The label indicates requires_api_key=false, meaning it does not depend on external API keys.

What Is This

weekly-report is a community-sourced (community) office efficiency Skill, currently at version 1.0.0, maintained by the publisher user_1f530050, and categorized as office-efficiency.

In one sentence: it is automatically triggered when a user expresses intents like “write a weekly report,” “generate a weekly report,” or similar phrasing. It pulls Git commits from all projects within a specified directory, analyzes them, generates a weekly report document, and saves it to a designated directory.

Official entry points:

  • SkillHub display page: https://www.skillhub.cn/skills/user_1f530050/weekly-report-new
  • API coordinate page: https://api.skillhub.cn/user_1f530050/weekly-report-new

Core Features

The following capabilities are drawn from the official introduction; configuration items not separately detailed on the page are not speculated upon:

  1. Intent Triggering: Recognizes expressions such as “write weekly report / generate weekly report / compile weekly report,” after which the Agent loads this Skill and executes the workflow.
  2. Multi-Project Git Aggregation: Within the directory specified by the user, it traverses the projects contained there and retrieves Git commit records (specific branches and time ranges are determined by the convention agreed upon with the Skill during the conversation; the page does not provide fixed configuration file fields).
  3. Analysis and Drafting: Inductively summarizes the commit records and outputs a formatted weekly report body.
  4. File Writing: Saves the generated weekly report to the directory specified by the user.

Installation and Activation

Installing SkillHub CLI (if not already installed)

First check whether the CLI is already present:

command -v skillhub && skillhub --version

If not installed, you can use the official one-click script (full installation includes default Skills; add --cli-only if only the CLI is needed):

curl -fsSL https://skillhub-1388575217.cos.ap-guangzhou.myqcloud.com/install/install.sh | bash

Installing weekly-report

The official page requires installation according to the SkillHub installation instructions, using the coordinate @user_1f530050/weekly-report-new. During installation, the --dir flag must point to the skills directory actually read by the current Agent; otherwise, the default landing in ./skills/ may not be recognized.

skillhub install @user_1f530050/weekly-report-new --dir <skills directory>

Commonly used skills paths for various clients (excerpted from the SkillHub installation documentation; select one according to your current environment):

Environment skills directory
Cursor ~/.cursor/skills/
Claude Code ~/.claude/skills/
Codex ~/.codex/skills/ or .agents/skills/ under the project
Gemini CLI ~/.gemini/skills/

After installation, if the Agent does not immediately load the new Skill, follow your client’s instructions to restart or refresh the skills cache.

The differences among Cursor, Codex CLI, and Claude Code mainly lie in the skills directory path and whether a session restart is needed; the installation command itself is skillhub install … --dir <path> in all cases, and the page does not provide a second, client-specific installation command.

Typical Usage

The official documentation does not provide an independent README-style step-by-step screenshot guide. A reproducible usage flow is as follows:

  1. Complete the installation described in the previous section, and confirm that the Agent can discover weekly-report.
  2. In the conversation, specify which projects to scan (parent directory path) and the weekly report output directory; if a time range needs to be constrained, state it clearly in the request (e.g., “this week,” “past 7 days”) to avoid the Skill guessing the interval on its own.
  3. Use a trigger phrase to initiate the task, for example:

Help me generate a weekly report: scan all git repositories under ~/work/repos, organize this week’s commits into a weekly report, and save it to ~/work/reports/.

  1. Check the generated weekly report file in the output directory, fine-tune the wording as needed, and submit it.

Prerequisites: The sub-projects under the target directory must be Git repositories, and git must be executable locally to read history; directories with no commits or no repositories will not produce valid content.

Applicable Scenarios and Notes

Who Is It For

  • Those with multiple Git projects mounted under a local or monorepo parent directory, where the weekly report needs to summarize “what was done” rather than serving as a single-repo changelog.
  • Those already using clients such as Cursor, Claude Code, or Codex that support Agent Skills, and who want to replace handwritten templates with natural language.

Limitations and Notes

  • The page does not declare how it handles edge cases such as non-Git directories, submodules, or shallow clones; for complex topologies, it is recommended to run a trial on a small number of repositories first.
  • The quality of “analysis” depends on whether commit messages are well-formed; empty messages or vague descriptions will affect the readability of the weekly report.
  • No API Key is required, but it depends on local Git and file system permissions; do not include directories to which you lack access in the scan path.
  • SkillHub statistics (based on captured data): rating 100000, approximately 810 downloads, approximately 3 installations; this is a community Skill, and upgrades and issue reporting should follow the SkillHub page and the publisher.

Conclusion

weekly-report consolidates the workflow of “multi-repository git log → formatted weekly report file” into a single Agent pipeline, making it suitable for scenarios where you regularly pull commits from a specified directory each week, produce a draft, and then save it. Installation and coordinates follow the official source:

https://www.skillhub.cn/skills/user_1f530050/weekly-report-new