ai_analysis
Puts a question to a model over the collected repository data and returns its analysis.
LABSCO SUMMARY
Reporting is separated from collection, so a report is a read of stored history rather than a fresh crawl of GitHub.
get_repo_data returns history for a repository out of the server's own data file, and the daily and weekly reports run over the repositories named in configuration rather than ones passed per call — which means the set being reported on is a deployment decision, not a prompt. Delivery is independent in the same way: ai_analysis answers a question about the stored data, and send_feishu_message will post whatever message you hand it, report or not.
What it is
A GitHub repository activity reporter with 5 tools: it keeps historical data for the repositories you configure, produces daily and weekly reports from it, runs AI analysis over the same data, and delivers messages to a Feishu chat.
What you get
- get_repo_data returns the stored historical data for a named GitHub repository.
- generate_daily_report and generate_weekly_report build the activity report for today and for the current week across the tracked repositories.
- ai_analysis puts a question to a model over the collected repository data and returns its analysis.
- send_feishu_message posts a message to the configured Feishu group.
Requirements
A GitHub token in GITHUB_TOKEN, the repositories to track in REPORT_REPOS, and a data file at DATA_FILE for the collected history. ai_analysis needs a model endpoint through API_KEY, API_BASE_URL and API_MODEL; send_feishu_message needs an incoming webhook in FEISHU_WEBHOOK.
Setup effort
One command plus a key — npx -y star-report mcp-server, then supply credentials
USE CASES
01
Posting a daily activity digest into the team chat
The report is built from stored history and the delivery is a separate call, so the same digest can go out on a schedule without a fresh crawl.
02
Summing up a week across the repositories you watch
The weekly report covers the configured set; pulling one repository's stored history alongside it is how you check a number that looks wrong.
03
Asking about the trend rather than reading the numbers
The analysis runs over the same collected data the reports use, so a question about direction does not need the raw series pasted in.
5 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
generate_daily_report
Generates today's activity report across the tracked GitHub repositories.
generate_weekly_report
Generates this week's activity report across the tracked GitHub repositories.
get_repo_data
Returns the stored historical data for a named GitHub repository.
send_feishu_message
Sends a message to the configured Feishu group.
BEFORE YOU INSTALL
What to know first
Tools you get
5 tools when we connected.
Requires
- npx on PATH
- GITHUB_TOKEN in your environment
- FEISHU_WEBHOOK in your environment
- API_KEY in your environment
CONTEXT COST
Light Standing cost
Rated by the size of the 5 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Add the block to your client, or run the command directly.
Server setup
Configuration file
JSON
{
"mcpServers": {
"starReport-server": {
"command": "npx",
"args": [
"-y",
"star-report",
"mcp-server"
],
"env": {
"GITHUB_TOKEN": "<YOUR_GITHUB_TOKEN>",
"FEISHU_WEBHOOK": "<YOUR_FEISHU_WEBHOOK>",
"REPORT_REPOS": "<YOUR_REPORT_REPOS>",
"API_KEY": "<YOUR_API_KEY>",
"API_BASE_URL": "<YOUR_API_BASE_URL>",
"API_MODEL": "<YOUR_API_MODEL>",
"DATA_FILE": "<YOUR_DATA_FILE>"
}
}
}
}npx
TERMINAL
npx -y star-report mcp-server
