Worklog Skill | Agent Skills

Worklog

个人工作回顾，整合本地 git 统计、GitHub 和云效数据，输出结构化 Markdown 报告。

Prerequisites

| Tool | Type | Required | Install | |------|------|----------|---------| | git | cli | Yes | brew install git or git-scm.com | | gh | cli | No | brew install gh then gh auth login | | jq | cli | No | brew install jq | | yunxiao | skill | No | Included in npx skills add niracler/skill | | yunxiao MCP | mcp | No | See yunxiao skill for setup |

Do NOT proactively verify these tools on skill load. If a command fails due to a missing tool, directly guide the user through installation and configuration step by step.

Mode Selection

| User says | Mode | |-----------|------| | 「日报」「今天做了什么」"daily worklog" | Daily | | 「周报」「本周总结」"weekly worklog" | Weekly | | 「工作回顾」「worklog」 | Ask user |

Workflow

Determine mode + dates → 2. Collect data (parallel) → 3. Integrate + render

If no data from any source, output: "No activity recorded for {period}."

Step 1: Determine Dates

| Mode | --since | --until | |------|---------|---------| | Daily | Today (YYYY-MM-DD) | Today | | Weekly | Monday of current week | Today |

Step 2: Collect Data (Parallel)

Run all three sources concurrently using subagents:

Other agent environments: the three sources are independent — run them sequentially if parallel execution is not available.

┌─ Bash: stats.sh --since {since} --until {until}       (local git)
├─ Bash: github.sh --since {since} --until {until}      (GitHub)
└─ subagent: 云效 MR/Bug/Task                            (yunxiao skill)

2a. Local Git Statistics

bash scripts/stats.sh --since YYYY-MM-DD --until YYYY-MM-DD

Output: JSON with period, repos[] (name, path, remote_url, commits, insertions, deletions, files_changed, authors, first_commit, last_commit), and totals (repos_active, repos_scanned, commits, insertions, deletions).

Other agent environments: run this command directly via shell.

2b. GitHub Activity

bash scripts/github.sh --since YYYY-MM-DD --until YYYY-MM-DD

Output: JSON with period, prs[] (repo, number, title, state, created_at, merged_at), issues[] (repo, number, title, state, created_at), and totals (prs_merged, prs_open, issues_opened, issues_closed).

Username is auto-detected via gh api user. No manual configuration needed.

Other agent environments: run this command directly via shell.

2c. 云效 Activity

Start subagent to invoke yunxiao skill with specific MCP queries:

Task tool:
- subagent_type: general-purpose
- prompt: 使用 yunxiao skill 获取 {since} 到 {until} 的工作记录

  先通过 MCP 自动获取身份：
  - get_current_user → 用户名
  - get_current_organization_info → 组织 ID

  然后查询：
  1. MR: list_change_requests — state=all, authorCodeupIdList=[user_id]
     返回: id, title, state, sourceRef, targetRef, createTime, mergeTime
  2. Bug: search_workitems — category=Bug, space=..., conditions=[assignee=username]
     返回: identifier, subject, status, assignedTo, gmtModified
  3. 任务: search_workitems — category=Task, space=..., conditions=[assignee=username]
     返回: identifier, subject, status, assignedTo, gmtModified

Other agent environments: if yunxiao skill is unavailable, skip and note "云效数据未获取".

Step 3: Integrate and Render

Merge all sources. For any failed source, note it (e.g., "GitHub 数据未获取") and continue with available data.

Render using template.md.

Weekly mode additionally requires:

Commit distribution — For active repos from stats.sh, count commits per weekday:

git log --since={since} --until={until_next_day} --author="{author}" --format="%aI" | \
  awk -F'T' '{print $1}' | sort | uniq -c

Render as ASCII bar chart (Mon–Fri).

Highlights — From stats.sh data: most active repo (by commits), largest change (by insertions).

Data Source Degradation

| Source failed | Behavior | |---------------|----------| | GitHub | Note "GitHub 数据未获取", continue | | 云效 | Note "云效数据未获取", continue | | Both remote | Output local git stats only |

Common Issues

| Issue | Fix | |-------|-----| | stats.sh empty repos | Check date range and --author filter | | github.sh fails / 401 | Run gh auth login | | 云效 MCP not connected | Check MCP server configuration |

Identity Detection

All user identities are auto-detected at runtime. See user-config.md for detection sources and prerequisites.

Agent Skills: Worklog

Install this agent skill to your local

Skill Files