Communitygithub.com

Agentchengfeng/chengfeng-videocut-skills

Organizes, sanitizes, and submits GitHub bugs for chengfeng-videocut. Use when the user says to report a bug, give feedback on video trimming issues, submit a GitHub Issue, tell the developer, or continue submitting a previously previewed bug draft. Do not use for feature suggestions, general troubleshooting, code commits, or auto-submission without user confirmation.

What is chengfeng-videocut-skills?

This skill provides a structured workflow for submitting GitHub Issues related to the chengfeng-videocut product. It guides the agent through verifying the bug is reproducible, selecting the correct repository (product vs. skills), generating a sanitized JSON input with minimal diagnostics, previewing a de-identified draft with the user, obtaining explicit confirmation, and finally submitting via a script. It enforces strict rules: no runtime installation, no media uploads, no code commits, and no sensitive data inclusion (API keys, tokens, personal info). The skill supports multiple platforms including Claude Code, Cursor, and GitHub Copilot via its SKILL.md conventions. It uses a two-step script workflow—first to generate and preview a draft with a confirmation token, then to submit only on user approval. This prevents accidental issue creation and ensures only verified, public-ready information is posted.

Works with✓Claude Code✓Codex CLI~Cursor
npx skills add https://github.com/Agentchengfeng/chengfeng-videocut-skills/tree/HEAD/plugins/chengfeng-videocut/skills/chengfeng-report-bug

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

Report a Bug

This is the support entry point for chengfeng-videocut, not a third editing workflow. By default, only a public Issue draft is generated; external submissions must go through sanitization preview and user confirmation.

Problem description
   |
   v
Minimal read-only diagnostics -> Sanitize -> Show repo / title / body
                                  |
                            User explicit confirmation
                                  |
                                  v
                     Authentication + label check + duplicate check
                                  |
                                  v
                         Create GitHub Issue

1. Confirm This Is a Reproducible Bug

First, distinguish:

  • Bug: Actual behavior deviates from existing contract, UI, version capability, or deterministic result.
  • Feature suggestion: The product never promised this capability; do not disguise as a bug.
  • Usage question: Answer or diagnose first, do not auto-submit.

Collect only the minimum facts needed for reproduction. Do not install Runtime, start Studio, modify projects, re-edit videos, or upload media to "gather evidence." If Runtime exists, you may run node scripts/ensure-runtime.cjs --json from the plugin root for read-only detection, keeping only state / kind / version / healthy / error.code; do not include the full doctor output. kind=desktop-managed only indicates the installation source is the desktop App, it does not authorize this Skill to launch the App, service, or reinstall. If it doesn't exist, write "not detected."

2. Select the Fixed Repository

The workspace root directory is not necessarily a Git repository; do not guess the remote from the current directory:

Issue ComponenttargetGitHub Repository
Runtime, Studio, UI, CLI, API, Timeline, RenderingproductAgentchengfeng/chengfeng-videocut
Skill, Marketplace, Installation Orchestration, MCP App, Workflow DocumentationskillsAgentchengfeng/chengfeng-videocut-skills

For cross-component issues, prioritize reporting to the repository that first violated the contract, and mention the other component in the body; if uncertain, ask the user first.

3. Generate Structured Input

Write the candidate as a JSON to the current task's temporary directory with umask 077, not to product specification files; delete the temporary JSON and Markdown after submission:

{
  "schemaVersion": 1,
  "target": "product",
  "component": "Studio",
  "title": "Runtime 0.1.1 cannot open HyperFrames top-level video trim view",
  "summary": "One-sentence impact description",
  "steps": ["Step one", "Step two"],
  "expected": "Expected behavior",
  "actual": "Actual behavior",
  "environment": {
    "Runtime": "0.1.1",
    "OS": "macOS"
  },
  "evidence": ["Only minimal, public diagnostics"],
  "acceptance": ["Verifiable fix criteria"]
}

Prohibited content:

  • API Key, Token, Cookie, .env;
  • Videos, audio, screenshots, full logs, or transcript bodies;
  • Client names, original video names, real projectId, user absolute paths;
  • Private repository content without user permission.

4. First Generate and Display the Sanitized Draft

Locate the script from the current SKILL.md:

First, locate the plugin root following the "Readiness Check" first step of Check for Updates (from the actual source file path of this Skill, two levels up; do not run codex plugin list or search other caches; substitute <plugin-root> below with that literal path). This Skill only uses it for script location, does not install Runtime or start services:

node "<plugin-root>/skills/chengfeng-report-bug/scripts/report-bug.cjs" draft --input "<report JSON file>" --output "<draft output file>" --json

The target must be explicitly product or skills; missing or misspelled values should stop, not silently redirect to another repository. The script cleans common secrets, local usernames, volume names, path fragments, and localhost query parameters; automatic sanitization cannot identify arbitrary client names, so manual review is still required. Display to the user:

repo
label=bug
title
Full sanitized body

Also save the returned confirmationReceipt, and show the user the confirmationToken and expiration time. Stop if the user does not explicitly confirm this public content; detecting an error does not equal consent to submit. The token is bound only to the draft content and does not itself prove user consent; the Agent must still wait for an explicit reply. Confirmation is valid only for the immediately following submission in the current task, expires in 30 minutes, and cannot be reused across tasks or sessions.

5. Submit After Confirmation

Submit only the draft just presented with unchanged content:

node "<plugin-root>/skills/chengfeng-report-bug/scripts/report-bug.cjs" submit --input "<report JSON file>" --target "<target repository>" --confirmed --confirm-token "<confirmation token>" --receipt "<confirmation receipt file>" --json

Fixed rules:

  • Before submission, the script checks gh auth status, the target repository's Issues, and the bug label; keep the draft on failure.
  • If a fingerprint matching the same component, reproduction steps, and actual results already exists, return the original Issue URL; do not impersonate a duplicate with a generic title.
  • If the body changes after confirmation, return confirmation_mismatch; must re-preview.
  • If the confirmation credential has expired or been used, return confirmation_expired / confirmation_replayed; must re-preview.
  • If the creation result has no explicit Issue URL, do not auto-retry; it may have already been created.
  • Only when you have the GitHub Issue URL can you say "the bug has been reported."

Failure Handling

  • github_auth_required: Indicates the draft is complete but not submitted; disallow claiming success.
  • github_issue_create_failed: Keep the draft and manual entry point; do not loop retry.
  • confirmation_required / confirmation_mismatch: Re-display the exact public content.
  • No permissions or visibility for the target repository: Do not revert to another repository; let the user decide whether to report to the public Skills repository instead.

Individual skills in this repo

This repo contains 7 individual skills — each has its own dedicated page.

Agentchengfeng/chengfeng-videocut-skills

剪辑环境的唯一管理者:就绪检查(skills 是否最新 → Runtime 是否配套)、Skills 更新激活、Runtime 安装与体检。用户说检查更新、安装剪辑环境、装播放器、检查剪辑环境、剪辑环境就绪了吗、配置转录凭证时使用;业务 Skill(剪口播/字幕/画面/导出)第 0 步也引用本 Skill 的就绪检查。不用于剪辑、字幕、画面、导出本身或项目数据迁移。

Agentchengfeng/chengfeng-videocut-skills

剪辑中文口播原素材:逐词转录、词典修字出修字表、五轮扫描找口误与重复、汇总表与重复句子表、打开 Studio 让用户复核、复盘沉淀用户偏好与词典。只产出一份已复核的删词账本,不切媒体、不做字幕、不做分镜动画。用户说剪口播、处理口误、生成口播基础素材、继续剪口播,或确认卡回传 action=return_cut_review 时使用。不要用于执行物理剪切、导出剪后视频、单独安装、单独打开工作台或口播分镜成片。

Agentchengfeng/chengfeng-videocut-skills

把剪好的口播烧成一个成片文件:账本切片段、推近、字幕、HTML 画面层,一次全部烧进 mp4。用户说导出、出成片、烧字幕、渲染、导出视频、生成最终文件时使用。不要用于生成删词候选、写字幕、做画面动画。

Agentchengfeng/chengfeng-videocut-skills

给剪好的口播做字幕:直接用已有的逐词稿加账本算出剪后时间(不必导出、不必重新转录)、用词典和作者文稿改写听错的专名、按句子分屏、在 Studio 里逐屏复核。用户说做字幕、加字幕、改字幕、重新分屏、字幕不对时使用。不要用于删词剪辑、物理剪切、分镜动画或成片渲染。

Agentchengfeng/chengfeng-videocut-skills

操作 chengfeng-videocut 剪辑工作台的已有工程,通过 Runtime CLI 查工程与词句位置、添加或替换画面覆盖层、移除画面、删除明确指定的词句或片段、拆分裁剪移动已有片段、调整画布比例并回读核验。用户说插入片段、删掉这句、移除动画、换个画面、调整顺序或管理时间线时使用;先区分画面覆盖与主轨顺延插入,不负责自动判断口误、创作动画、制作字幕、导出或排查安装故障。

Agentchengfeng/chengfeng-videocut-skills

Create Ian Xiaohei-style HTML/SVG motion illustrations for Chinese articles, scripts, storyboard scenes, workflow explanations, concept metaphors, and short video visual aids. Use when the user asks for 小黑 SVG、漫画感 HTML、分层 SVG 动效、把小黑生图做成 HTML/SVG、可编辑矢量动效、口播流程动画、手绘漫画动效、正文配图动画, or wants a Xiaohei illustration style adapted into controllable HTML/SVG/GSAP output instead of raster image generation.

Agentchengfeng/chengfeng-videocut-skills

给剪好的口播配画面:在录屏上盖 HTML 层(圈重点标注 / 小黑整屏动画 / 推近),层绑字幕屏、由播放器逐帧驱动、直接在预览里看。用户说做分镜、配画面、加动画、圈重点、B-roll、做 storyboard 时使用。不要用于删词剪辑、字幕、物理剪切或成片渲染。

Related Skills