Audio2Text

将音频转文字，并提供 AI 总结、录音要点与发言人区分。 音频转文字：云端转写，支持多格式（mp3/wav/m4a 等）。 AI 总结：自动生成摘要与纪要。 录音要点：章节、待办、推荐等结构化要点。 发言人区分：转写结果带发言人标识，便于会议/对话场景。

Security Scan

The skill appears to do what it says (upload a local audio file to Tinrec, request transcription/summarization, and return structured results); code and instructions are consistent with a cloud-based transcription flow, but there are a few small inconsistencies and privacy/operational concerns you should be aware of before installing or using it.

Purpose & Capability

Name/description (audio→text, summaries, speaker separation) align with the code and SKILL.md: the CLI uploads a local audio file to a Tinrec API, submits a transcription job, polls status, and returns transcript/summary/speaker info. No unexpected third-party services beyond the stated Tinrec endpoints are used.

Instruction Scope

SKILL.md and the CLI instruct the agent/operator to read a local audio file and to persist a user-provided API key in an 'api-keys' file (or provide via env/CLI). The skill will upload the audio and metadata to the Tinrec service (external network transmission) — this is expected for a cloud transcription service, but it is a privacy-sensitive operation and the instructions explicitly require writing/storing API keys locally.

Install Mechanism

No install/spec—instruction-only plus a single Python script that uses only the standard library (urllib/json). Nothing is downloaded or installed by the skill itself; there is no extract or external package install step.

Credentials

Registry metadata lists no required env vars, but the CLI reads an API key from: (1) the --api-key argument, (2) environment variable TINREC_API_KEY, or (3) an api-keys file. The skill therefore effectively requires a service credential, but the manifest did not declare that. This mismatch (undeclared required secret) is an inconsistency to note. The number and scope of environment accesses themselves are otherwise proportionate to the stated purpose (one service API key).

Persistence & Privilege

always:false and no system-wide modifications. The skill asks the operator/agent to save an API key to a local file if desired, but it does not request persistent platform privileges or modify other skills or system configurations.

Assessment

This skill is coherent with its stated purpose (cloud transcription): it uploads a local audio file to a Tinrec endpoint, submits a transcription job, polls for completion, and returns transcript, summary, chapters, todos, and speaker data. Before installing/using it consider: 1) Privacy: your audio is uploaded to the Tinrec service (default base URL https://api.tinrec.com/api). Only proceed if you trust that service and its privacy/retention policies. 2) API key handling: the SKILL.md recommends saving user-provided API keys into a local 'api-keys' file; the script also accepts TINREC_API_KEY. Storing secrets on disk can be risky—prefer using short-lived or scoped keys, and avoid leaving keys in plaintext if possible. 3) Manifest mismatch: the skill metadata lists no required env vars, but the code requires an API key; treat that as a minor inconsistency and ensure you supply a valid key via CLI/env when using. 4) Source/trust: the registry entry has no homepage and the owner is an opaque ID — if you need stronger assurance, obtain the script from a known repository or contact the publisher for provenance. If you accept the privacy implications and provide a dedicated API key, the skill appears suitable for its stated use. If you cannot trust the destination service or cannot risk uploading audio, do not use this skill.

使用说明

使用tinrec的云端API进行便捷的音频转文字，区分发言人并快速总结音频内容和生成摘要

使用流程

1. 先获取api-keys，获取地址https://tinrec.com/api-keys，注册账号后即可免费获取 Key 1.1. claw需要将用户发送的key保存至api-keys文件后使用cli模式进行调用，调用时通过 --api-keys-file 指定该文件，每位用户注册后可免费领取1小时转写额度，开通Pro后解锁转写限制

在技能目录下使用 Python CLI，一次完成上传、转写、轮询并输出结果：

cd script
# 使用目录下的 api-keys 文件
python audio2text_cli.py /path/to/audio.mp3 --api-keys-file ./api-keys

• --api-keys-file：API Key 文件路径（默认当前目录 api-keys），文件内第一行非空非注释即为 Key。
• --api-key：直接传入 Key，优先于文件和环境变量。
• --base-url：API 地址，默认 https://api.tinrec.com/api。
• --json：仅输出 JSON（含转写、总结、要点、发言人），便于 AI 解析。
• --no-wait：只提交不轮询，返回任务 id。
• --poll-interval / --timeout：轮询间隔与总超时。

输出包含：summary（AI 总结）、chapters/todos（要点）、transcripts（转写正文）、speaker_data（发言人信息）。
状态说明：接口有两个状态——转写状态（transcript_status）与任务状态（task_status）。转写完成后才有正文；任务状态完成后才会有总结、要点、发言人等信息，CLI 会轮询到两者均完成后再输出。

使用场景

在以下场景下应调用本技能（传入用户提供的本地音频文件路径及可选的 --api-keys-file）：

• 用户发来/提到一段本地录音文件，希望转成文字、要全文或摘要时。
• 用户明确要求「把这段音频转成文字」「听写/转写这段录音」，且能访问到本地文件路径时。
• 用户需要会议/访谈纪要：要总结、要点、待办或「谁说了什么」的发言人区分时。
• 用户提供录音并希望得到结构化结果：如章节、待办列表、推荐后续动作等，且文件在本地可读时。
• 用户说「用 Tinrec / 用你们的转写」 或指定要用该服务处理本地音频时。

不适用：仅提供在线链接/URL 而无法下载为本地文件时，需先下载再调用；无本地文件路径或用户未提供音频文件时不调用。