Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/derek-larson14/delegatenpx agentmods add commands/derek-larson14/delegate/setup-transcriptionWrote 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/commands/derek-larson14/delegate/setup-transcription)<a href="https://agentmods.dev/commands/derek-larson14/delegate/setup-transcription"><img src="https://agentmods.dev/badge/commands/derek-larson14/delegate/setup-transcription/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/commands/derek-larson14/delegate/setup-transcription"><img src="https://agentmods.dev/badge/commands/derek-larson14/delegate/setup-transcription.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 | $0.00017 | $0.07752 |
| Opus 5.5 | $0.00007 | $0.03101 |
| Sonnet 5.5 | $0.00003 | $0.01550 |
| Haiku 4.5 | $0.00002 | $0.00775 |
Grade C, and why
setup-transcription scanned grade C with 2 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 1mo 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf /tmp/hear /tmp/hear.zip Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -sL https://sveinbjorn.org/files/software/hear.zip -o /tmp/hear.zip && \ How it starts
The opening of the file, as written. The whole thing — 861 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Voice Transcription
Transcribe voice recordings and add them to voice.md. Supports two sources: iPhone Voice Memos (via iCloud, Mac only) and Google Drive (via Pigeon app, all platforms).
Step 0: Detect OS
uname -s 2>/dev/null || echo "Windows"
If the output contains "MINGW", "CYGWIN", "MSYS", or "Windows", this is a Windows environment. On Windows:
- Voice Memos source is not available (iCloud/Voice Memos are Mac-only)
- Auto-set source to "drive" (skip the source question)
- Use PowerShell scripts (.ps1) instead of bash (.sh)
- Use Task Scheduler instead of launchd
- No local
heartranscription — uses companion.mdtranscripts from Apps Script or Pigeon on-device
Step 1: Determine source
Check if source is already configured:
mkdir -p .voice && ([ -f .voice/source ] && cat .voice/source || echo "NOT_SET")
If Windows: auto-set to "drive" without asking:
mkdir -p .voice && echo "drive" > .voice/source
Tell user: "On Windows, voice transcription uses Google Drive (Pigeon app). Voice Memos is Mac-only."
If macOS and "NOT_SET", use AskUserQuestion:
"Where are your voice recordings?"
Options:
- "Voice Memos (iPhone → iCloud → Mac)"
- "Google Drive (Pigeon app)"
- "Local folder (iCloud Drive, Dropbox, or any folder path)"
- "Mac Shortcut (record directly on your Mac)"
If "Mac Shortcut", share this link and set up the folder:
Tell the user: "Install the Pigeon Shortcut: https://www.icloud.com/shortcuts/c323898b57ec44689d670aa5775b0761 — it records audio, transcribes on-device, and saves a transcript. On first run, pick your transcripts folder (default: ~/pigeon/transcripts/)."
mkdir -p ~/pigeon/transcripts
echo "mac" > .voice/source
echo "$HOME/pigeon/transcripts" > .voice/local-path
Then skip to the auto-routing section — no scheduled transcription needed since the Shortcut handles it on demand.
Save their choice:
echo "voicememos" > .voice/source # or "drive" or "local"
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.
- 1mo ago First seen · 861 lines · 17 tokens per session scan C 850f3fac044c
setup-transcription is a command published in the GitHub repository derek-larson14/delegate (39 stars, last pushed 6mo ago), licensed MIT. It adds 17 tokens to every session and 7,752 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it C with 2 findings (recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
brand-generate
Generate an on-brand document from a saved Brand Profile.
cti-report
Render case deliverables — relationship graph (PNG/SVG/Mermaid) and a polished PDF/DOCX assessment. Usage: /cti-report [--graph|--pdf].
render-figures
Compile all .tex and .typ figure files in a directory.
ppt-image2-editable-rebuild
Rebuild image2 or imagegen reference slides as editable PowerPoint decks.
publish-api-docs
Generate interactive API documentation from blueprint specs (Swagger UI, Redoc, AsyncAPI).
write-patent
An automated workflow for creating a complete Chinese patent application from a technical disclosure document, including written sections, patent drawings, and a Word file.