Work Report Generator

The Problem It Solves

Writing work reports is a common pain point. Whether due to tight deadlines, scattered information, or the tendency to produce stiff, formulaic text that reads like AI-generated officialese, the process is often tedious. This skill addresses both issues: it structures raw work records into coherent daily, weekly, monthly, and annual reports, and it enforces strict "anti-AI tone" rules to ensure the output sounds natural and conversational, avoiding the artificial polish of typical corporate summaries.

Core Capabilities & Workflow

The skill is built around two core concepts: layered aggregation and casual input processing.

1. Hierarchical Generation with Automatic Linking
   The system follows a daily -> weekly -> monthly -> annual logic. When generating a higher-level report, it automatically and silently reads and synthesizes the corresponding lower-level historical files to ensure consistency and avoid repetition. All reports are stored in a fixed structure, e.g., daily reports at {workspace}/.workbuddy/reports/daily/2026-06-05.md.

2. Accepts Any Input Format & Auto-Identifies Report Type
   You don't need to write formal descriptions. Input can be a few keywords (e.g., "login bug, data export") or a casual sentence (e.g., "mainly finished the login optimization this week"). The skill will:

   • Detect Intent: Automatically determine whether to generate a daily, weekly, monthly, or annual report based on explicit commands ("write weekly report") or contextual keywords (containing "week").
   • Intelligent Expansion: Expand keywords or casual phrases into natural short sentences while strictly adhering to the anti-AI tone norms. For example, input "API integration got stuck" becomes "The API integration hit a snag, their interface kept erroring," not "Technical challenges were encountered during the interface integration process."

3. Enforced Natural Language (Anti-AI Tone)
   This is a mandatory constraint. After generation, the skill runs a self-check, banning AI-style patterns like "firstly...secondly...finally..." structures or overuse of the particle "了" (completed). It encourages conversational transitions (e.g., "but," "actually") and concise numeric descriptions. If the check fails, it regenerates until the text reads like it was written by a human colleague.

Applicability & Important Notes

• Use Cases: Ideal for developers, product managers, and project members who need to quickly organize daily work logs into structured reports. It’s especially suited for those who are comfortable speaking but dislike formal writing.
• Key Considerations:
  • No Fabrication: All content must originate from your input or existing historical reports. The skill does not invent data.
  • Word Count Control: It has default recommended word ranges (e.g., 150-300 words for daily reports) with adjustable options.
  • Conversational Tone First: Natural language is prioritized over "professionalism." Problem descriptions should be honest, avoiding vague phrases like "needs improvement."
  • History Dependency: Generating a weekly report relies on this week's daily reports; a monthly report relies on weekly reports. If history is empty, the skill will ask minimally but aims to minimize interruption.