speech-transcribe-local SKILL.md skillA

speech-transcribe-local from dengxianghua888-ops/reel-scribe is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add dengxianghua888-ops/reel-scribe/speech-transcribe-local@git:20260927.9df448d

Pins git:20260927.9df448d in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
dengxianghua888-ops/reel-scribe
Path
atomic/speech-transcribe-local/SKILL.md
Versions recorded
1
Latest version
git:20260927.9df448d, committed Sep 27, 2026
Size
2,160 bytes
Repository stars
4
License
none declared

使用已固定哈希的本地 faster-whisper 模型识别指定音轨与源时间窗口,输出原始识别、词时间与不确定项;不改原话、不删口癖、不挑片段、不辨认说话人。.

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

通用调用先按[能力交接合同](../../contracts/capability-handoff.md)核对当前 Agent 的读取、执行和读回能力。下文的操作名、严格字段、路径和本地执行器只约束选择本库实现的路径;使用其他工具时保留目标、范围、保护项与结果核对语义。

# 本地语音观察

按[本地ASR合同](../../contracts/local-asr-contract.md)调用`operation=speech-transcribe-local`。必须在已核验的Python依赖环境运行;模型、依赖清单、素材和音轨窗口均显式绑定,不隐式下载模型或上传媒体。语言和识别参数由调用者给定,原子不判断素材是否适合目标风格。

原始结果保存`raw-observation.json`。只有全部词的文本/非零时间/概率结构合法才生成`word-transcript/1`;任一词异常时保留`REVIEW_REQUIRED`和原始结果,不均分时间、不丢词再交付一份貌似完整的稿。`NO_WORDS_DETECTED`只是模型未识别出词,不证明没有语音。

模型目录可使用CT2文本或JSON词表;存在的预处理配置和所有词表必须进入模型锁。模型格式兼容不等于对当前语言/专名更准确;替换模型保留新旧观察及相同输入绑定,不覆盖原稿或自动取代原先选择。

`model-aligned`表示模型内部的时间观察,不是人工听检或独立强制对齐。`speaker=unknown`不得改为已知人物。模型置信度高也可能幻觉,未识别区间也不能自动当作静音;是否删减、纠错和选段由复杂层回源判断。需要不支持的云Provider、语言或说话人分离时明确报告缺项。

执行器:[asr_ops.py](../../runtime/asr_ops.py)。运行成功不表示识别准确或声学切点自然;下游先审查转写与边界,再调用[transcript-import](../transcript-import/SKILL.md)和[speech-plan-compile](../speech-plan-compile/SKILL.md)。

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add dengxianghua888-ops/reel-scribe/speech-transcribe-local@sha256:221d5ae1e9eeb5e1

The line at the top of this page pins the label git:20260927.9df448d, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_x7hod6obz35p73cb.svg)](https://markdownregistry.com/a/art_x7hod6obz35p73cb)

3 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 1 version of this file, committed on Sep 27, 2026.

versioncommittedcommitsizeaudit
git:20260927.9df448d latest2026-09-27 9df448d 2,160 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (2160 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

dengxianghua888-ops/reel-scribe · 4 stars · license none declared · pushed 2026-11-08 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_x7hod6obz35p73cb
GET https://markdownregistry.com/api/v1/resolve?ref=dengxianghua888-ops/reel-scribe/speech-transcribe-local
GET https://markdownregistry.com/api/v1/blob/221d5ae1e9eeb5e185ef2e80229d0559e2c34d36b3f41c79b7e4e0824cb008cf

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from dengxianghua888-ops/reel-scribe

AGENTS.md AGENTS.md file
dengxianghua888-ops/reel-scribe · AGENTS.md
git:20260927.45ff613 · audit A · 4 stars
audio-envelope SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/audio-envelope/SKILL.md · 按明确源区间、增益、采样率和淡入淡出长度处理现有音频,生成可放入作品的WAV;不选择环境声、不决定转场、不判断对白或自动降噪。
git:20260927.9df448d · audit A · 4 stars
audio-extract SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/audio-extract/SKILL.md · 按指定源时间范围、采样率和声道数提取 PCM WAV;用于转写或音频处理前准备,不选择台词、不清除停顿。
git:20260927.9df448d · audit A · 4 stars
audio-mix SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/audio-mix/SKILL.md · 将指定音频区间按明确时间位置和增益混入已有视频,保留原视频流;用于已确定的配音或配乐摆放,不选择音乐、不改写旁白、不自动裁短语音。
git:20260927.9df448d · audit A · 4 stars
caption-burn SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/caption-burn/SKILL.md · 按明确输出时间、字体文件、字号、颜色和边距烧录给定字幕;不改文案、不选择关键词、不自动设计字幕风格。
git:20260927.9df448d · audit A · 4 stars
caption-coverage-check SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/caption-coverage-check/SKILL.md · 按明确文字、字幕ID和成片时间窗核对当前作品字幕覆盖,保留中文、数字与标点;不判断主张真假、措辞充分性、听感或画面可读性。
git:20260927.9df448d · audit A · 4 stars
caption-map SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/caption-map/SKILL.md · 将给定原文字幕按确定剪辑计划从源时间映射到输出时间,生成 JSON 与 SRT;不转写、不校正文案、不选择强调词。
git:20260927.9df448d · audit A · 4 stars
color-lut-apply SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/color-lut-apply/SKILL.md · 对已声明BT.709 SDR色彩解释和零起点恒定帧率视频应用指定3D cube LUT,保留帧顺序和音轨;不选择风格、不自动匹配参考色调、不做HDR或Log转换。
git:20260927.9df448d · audit A · 4 stars
cue-alignment-check SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/cue-alignment-check/SKILL.md · 按已给音乐源时间标记、当前音轨和指定画面帧检查提前或延后量及容差,返回全部数值匹配结果;不识别节拍、乐段、动作或选择切点,不修改作品。
git:20260927.9df448d · audit A · 4 stars
event-plan-compile SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/event-plan-compile/SKILL.md · 校验上层指定事件的铺垫、动作、结果在作品中的完整映射与顺序,检查回放关联和可见标识并分别计数;不发现事件、不判断胜负、不自动识别语义重复。
git:20260927.9df448d · audit A · 4 stars
evidence-plan-compile SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/evidence-plan-compile/SKILL.md · 将上层明确指定的内容单元、来源区间、保留依赖和章节映射到当前作品,检查证据完整覆盖及先修顺序,支持独立录音;不挑重点、不判断建议是否通过、不认证教学完整性。
git:20260927.9df448d · audit A · 4 stars
frame-extract SKILL.md skill
dengxianghua888-ops/reel-scribe · atomic/frame-extract/SKILL.md · 按明确时间点从已登记视频抽取画面 PNG;用于建立观察证据,不自行寻找高光或判断最佳镜头。
git:20260927.9df448d · audit A · 4 stars

Every file in dengxianghua888-ops/reel-scribe

Browse by kind, by grade A, or by owner.