メニュー
Use when the main execution agent must produce a final Markdown delivery report after the task is complete, especially when reviewer status is done and the report must be saved to a specific local file for human/mobile reading and bot delivery.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | final-task-report |
| description | Use when the main execution agent must produce a final Markdown delivery report after the task is complete, especially when reviewer status is done and the report must be saved to a specific local file for human/mobile reading and bot delivery. |
Use this skill only in the final handoff phase.
Produce one Markdown report file that a user can read directly on phone or desktop.
The report must stay grounded in the actual work completed in the repository and current session. Do not invent experiments, datasets, or code changes.
This skill is meant to be updated over time as the reporting format evolves. Main agent should follow this skill directly for final report generation.
Always include these sections in order:
## 0. Original User Task or ## 0. 原始用户任务## 1. What This Task Proposed or ## 1. 本次实验提出了哪些点## 2. How It Was Done and What Data Was Used or ## 2. 具体是怎么做的，使用了什么数据## 3. What the Main Agent Changed, What Approach It Used, and What Experiments It Ran or ## 3. main agent 修改了什么，使用了什么思路，跑了什么实验## 4. Special Notes or ## 4. 特殊说明If the task context is mixed-language, bilingual section titles are acceptable.
未使用 or none.未运行额外实验 or none.Use this skill any time a .pptx file is involved — creating, reading, editing, or generating a run report presentation. This includes: creating slide decks for mentors/colleagues/classmates, generating automated run report presentations, reading or parsing .pptx files, editing existing presentations, and using PptxGenJS. Trigger whenever the user mentions 'deck', 'slides', 'presentation', 'PPT', or references a .pptx filename.
Detect when a BTW question is asking for project files, screenshots, or effect images, then return the most relevant local file paths quickly.
Operate Codex as a high-autonomy repository executor for implementation, debugging, and long-running research loops. Use when work requires strict git discipline (init, pull, commit, optional push), repeated continuation until the objective is fully complete, and strong runtime validation such as smoke runs (at least 100 steps), checkpoint integrity checks, and basic inference verification before stopping.
Use when a planning or manager agent must maintain architecture, framework decomposition, TODO tables, and next-session objectives while also exploring missing context from the repository, official docs, external APIs, datasets, or similar open-source projects before handing concrete work to an execution agent.