Getting it into your agent
There is no command for this one: it runs only inside a plugin, and the catalogue could not identify which plugin ships it. The source is linked below.
Wrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/chenyuxiaojin/xiaochen-skills/cyxj-subfix)<a href="https://agentmods.dev/skills/chenyuxiaojin/xiaochen-skills/cyxj-subfix"><img src="https://agentmods.dev/badge/skills/chenyuxiaojin/xiaochen-skills/cyxj-subfix/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/chenyuxiaojin/xiaochen-skills/cyxj-subfix"><img src="https://agentmods.dev/badge/skills/chenyuxiaojin/xiaochen-skills/cyxj-subfix.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00079 | $0.02870 |
| Opus 5 | $0.00039 | $0.01435 |
| Sonnet 5 | $0.00016 | $0.00574 |
| Haiku 4.5 | $0.00008 | $0.00287 |
Grade A, and why
cyxj-subfix scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 191 lines — stays where its author put it; the contents beside it link to each section on GitHub.
达芬奇字幕修正 Skill v4
路径约定
脚本一律用 ${CLAUDE_PLUGIN_ROOT}(Claude Code 加载插件时自动注入)定位:
python3 "${CLAUDE_PLUGIN_ROOT}/skills/cyxj-subfix/srt_cleaner.py" ...
架构原则
三层分离:Python 做结构处理,Gemini 做语义初修,Opus 审查把关。
- Python(srt_cleaner.py):HTML 清理、去重、标点替换、合并、拆分、编号、导出逐字稿 Markdown — 所有时间码操作
- Gemini(srt_corrector.py):同音字修正 + 去口吃 + 去填充词 — API 自动处理,几乎免费
- Opus(对话内):审查 Gemini 的修改清单,纠正错误修正,发现遗漏 — 只看 diff,token 极少
词典定位:词典是"已知陷阱"参考,不是主力。同一个词的错误变体太多,词典永远无法穷举。真正的修正靠 Gemini 的语义理解 + 视频主题上下文。Opus 审查后发现的遗漏会反馈到词典,让系统越来越准。
使用方法
/字幕修正 ~/Desktop/Timeline1.srt # 完整流程
/字幕修正 ~/Desktop/Timeline1.srt --topic "Claude Code教程" # 指定主题提高修正准确度
/字幕修正 ~/Desktop/Timeline1.srt --no-regroup # 跳过合并拆分
/字幕修正 ~/Desktop/Timeline1.srt --premium # Phase 2a 用高端 Gemini 模型
/字幕修正 添加词条 错误词→正确词 # 添加词典条目(见下方"词典管理")
显示规则
- 场景:横屏 16:9
- 软上限:18 字符(正常通过)
- 警告区:18-25 字符(输出时标注提醒)
- 硬上限:25 字符(Python 强制拆分,标记需人工校验)
- 字符计算:中文/中文标点=1, ASCII=0.5, 向上取整
- 标点规则:字幕中不用逗号句号(替换为空格),保留?!
执行步骤
Phase 1: Python 结构处理
- 备份:创建
.bak备份文件(保住原始 SRT——后续各 Phase 会在同目录产出多个文件,防误覆盖,也便于最后回滚比对) - 运行 srt_cleaner.py:
如果用户指定python3 "${CLAUDE_PLUGIN_ROOT}/skills/cyxj-subfix/srt_cleaner.py" "<input.srt>" --stats--no-regroup,追加该参数跳过合并拆分。 - 报告结果:向用户汇报去重、合并、拆分、超限条目数量
Phase 2a: Gemini 自动初修
- 确认主题:如果用户提供了
--topic,以此为语境参考;否则从字幕内容推断主题,向用户确认 - 运行 srt_corrector.py:
如果用户指定python3 "${CLAUDE_PLUGIN_ROOT}/skills/cyxj-subfix/srt_corrector.py" "<_cleaned.srt>" --topic "主题"--premium,追加该参数使用高端模型。 - 输出:
_gemini_fixed.srt:Gemini 修正后的完整 SRT_changes.json:修改清单(原文→修正,带原因)
Phase 2b: Opus 审查(在对话中)
- 读取
_changes.json(只看修改清单,不读全文 SRT) - 审查 Gemini 的修正:
- 检查每项修正是否合理
- 标记错误的修正(Gemini 改错了的)
- 发现遗漏的错误(浏览 _gemini_fixed.srt 中 Gemini 未修改的部分,抽查是否有遗漏)
- 应用最终修正:
- 撤销错误的修正
- 补充遗漏的修正
- 保存为
_fixed.srt
- 反馈到词典:
- 把 Gemini 漏掉的重要错误写入
dictionary.json的feedback.gemini_missed字段 - 如果遗漏的错误有通用性(不是偶发的),同时添加到
corrections中
- 把 Gemini 漏掉的重要错误写入
What ships with it
7 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 191 lines · 79 tokens per session scan A 505325d8a81b
cyxj-subfix is a skill published in the GitHub repository chenyuxiaojin/xiaochen-skills (6 stars, last pushed 7d ago), licensed MIT. It adds 79 tokens to every session and 2,870 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
canva-asset-manager
Upload and manage assets (images, videos, audio) in your Canva account. Use this skill when user wants to upload files to Canva, manage their asset library, or work with uploaded media. Handles image, video, and audio uploads. Follows 3-mode workflow. For organizing uploaded assets use canva-folder-organizer.
canva-manager
Comprehensive Canva API integration skill for managing designs, folders, assets, and exports. Uses a 3-mode workflow: PLAN (analyze what needs to be done), CLARIFY (ask user questions), and IMPLEMENT (execute changes via Python scripts). Ensures safe operations by always confirming changes before execution. Use this…
canva-brand-kit
Manage and apply brand kits in Canva. Use this skill when user wants to work with brand colors, fonts, logos, or maintain brand consistency across designs. Helps define brand guidelines and ensures designs follow brand standards. Requires Canva Pro/Enterprise for full brand kit features.
canva-content-generator
Generate content ideas and text for Canva designs. Use this skill when user needs help with what to write, content suggestions, copy for designs, captions, headlines, or creative direction. Analyzes user's sample preferences from the samples folder to match their style. Does NOT directly edit Canva - provides content…
canva-export
Export Canva designs to various formats (PDF, PNG, JPG, PPTX, MP4, GIF). Use this skill when user wants to download, export, or save their designs. Handles single and bulk exports with format options. Follows 3-mode workflow for confirmation. For editing designs use other specialized skills (canva-image-editor…
canva-folder-organizer
Organize and manage folders in your Canva account. Use this skill when user wants to create folders, move designs between folders, rename folders, or organize their Canva project structure. Follows 3-mode workflow to prevent accidental moves or deletions. For viewing folder contents use canva-explorer first.