Project Documentation

Cloud Sync Workflow

由项目内 Markdown 源文件转换生成的 HTML 版本,用于浏览器阅读和团队共享。

Cloud Sync Workflow

This workflow standardizes how everyone uses the project:

  1. Pull the latest cloud version.
  2. Run the task.
  3. Commit and push the result back to the cloud.

The executable entrypoint is:

bash scripts/sync_workflow.sh

First-Time Setup

Beginners should not run Git commands manually. Send them docs/codex-first-use.md; Codex will check Git capability, clone or update the repository, read AGENTS.md, and report any Codeup authorization blocker.

After Codex has pulled the project locally, it can check that the cloud remote is visible:

bash scripts/sync_workflow.sh status

If Git asks who you are, configure your identity once:

git config user.name "Your Name"
git config user.email "you@example.com"

Daily Beginner Flow

Before doing anything:

bash scripts/sync_workflow.sh pull

After editing files or generating a report:

bash scripts/sync_workflow.sh upload "sync: update bid analysis"

To upload only selected files or directories:

bash scripts/sync_workflow.sh upload "sync: update docs" -- README.md workflows/

If one command performs the whole task, use run so pull and upload happen automatically:

bash scripts/sync_workflow.sh run "sync: profile bid data" -- python3 scripts/profile_bid_data.py data/raw/your-file.csv

Agent Flow

When an AI assistant works in this repository, it should follow AGENTS.md:

  1. Run pull before answering or editing.
  2. Make the requested changes.
  3. Run upload before the final answer.

This is the project-level rule that turns each collaboration round into a cloud synced Git workflow.

What Gets Uploaded

The upload command runs git add -A, so normal repository files are committed. The existing .gitignore still applies:

Use this rule to avoid accidentally publishing sensitive tender data, contacts, phone numbers, or customer clues.

Failure Handling