Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Soboljem/transcribe-mcp --skill dead-air-cuttergit clone --depth 1 https://github.com/Soboljem/transcribe-mcpWrote 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/soboljem/transcribe-mcp/dead-air-cutter)<a href="https://agentmods.dev/skills/soboljem/transcribe-mcp/dead-air-cutter"><img src="https://agentmods.dev/badge/skills/soboljem/transcribe-mcp/dead-air-cutter/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/soboljem/transcribe-mcp/dead-air-cutter"><img src="https://agentmods.dev/badge/skills/soboljem/transcribe-mcp/dead-air-cutter.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 | $0.00108 | $0.01747 |
| Opus 5.5 | $0.00043 | $0.00699 |
| Sonnet 5.5 | $0.00022 | $0.00349 |
| Haiku 4.5 | $0.00011 | $0.00175 |
Grade A, and why
dead-air-cutter 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 22d 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 — 111 lines — stays where its author put it; the contents beside it link to each section on GitHub.
dead-air-cutter
Turns talk.mp4 into talk_cut.mp4 and talk_cut.srt. The pauses are gone, and the subtitles
line up with the cut version, not the original.
What it needs — check before starting
- ffmpeg and ffprobe on PATH (
ffmpeg -version). If they're missing, tell the user and stop. - Node 20+ (
node --version). It's already there ifnpxruns MCP servers. - The
transcribe-mcpMCP server, version 0.3.0 or newer. It has to be configured withTRANSCRIBE_MCP_WALLET_KEY: the private key of a dedicated wallet holding a few dollars of USDC on Base. Callshop_infofirst. It's free and says whether the wallet is ready. Iftranscribe_urlhas nofileparameter, the server is older than 0.3.0. Tell the user to update it: puttranscribe-mcp@latestin the MCP config and restart the client.
What leaves the machine — tell the user once, up front
- The audio only (never the video) goes to
https://transcribe.soboljem.cz, compressed to mono Opus. It's transcribed there by Whisper on Cloudflare Workers AI. - Each call is paid from the user's wallet in USDC over x402. That's $0.05 per ≤6-minute part, plus $0.002 for the cut list. A 10-minute video costs about $0.10, and an hour about $0.50. The MCP server refuses any single quote above $0.10.
- For the cut list, only the timings (start/end of each speech segment) are sent. Nothing else leaves the machine.
- The shop never stores the audio. It keeps a copy of the result for up to ~8 hours. That copy is a shop-side safeguard this client can't use: calling again always pays again. Each sale also leaves a permanent row in the shop's books: wallet address, tx hash and counts, the same things that are public on Base. Payment goes through the PayAI facilitator. The full list is at https://transcribe.soboljem.cz/terms. Point the user there if they ask.
What runs locally
scripts/dac.mjsrunsffmpegandffprobe(throughexecFileSync, no shell). It reads the video. It writes into the work folder (audio parts,manifest.json, transcripts,segments.json,deadair.json,cuts.json), and writes<name>_cut.mp4and<name>_cut.srtnext to the video. A previous cut with the same name is overwritten. On a fresh split it deletes its own leftovers in the work folder (part*.opus,manifest.json,segments.json,deadair.json,filter.txt), but never a paid transcript orcuts.json.- It makes no network calls and never touches the wallet. Only the MCP server talks to the shop and signs payments. This skill never reads the private key.
What ships with it
3 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.
- 22d ago First seen · 111 lines · 108 tokens per session scan A af6c75d77ab0
dead-air-cutter is a skill published in the GitHub repository Soboljem/transcribe-mcp (0 stars, last pushed 23d ago), licensed MIT. It adds 108 tokens to every session and 1,747 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-10-17.
dead-air-cutter compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| dead-air-cutterSoboljem/transcribe-mcp | Skill | 0 | 23d ago | MIT | A | 108 | Claude Code, Codex |
| imagegenopenai/codex | Skill | 134k | today | Apache-2.0 | A | 113 | Codex |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | yesterday | MIT | A | 200 | Claude Code, Codex |
| playwright-climicrosoft/playwright | Skill | 100k | yesterday | Apache-2.0 | A | 19 | Claude Code |
| browser-automationopenclaw/openclaw | Skill | 394k | today | MIT | A | 36 | Claude Code |
| systematic-debuggingobra/superpowers | Skill | 315k | 5d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | today | MIT | A | 95 | Claude Code |