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 agentmods add instructions/alexanderabramovpav/scriba/agents-mdgit clone --depth 1 https://github.com/AlexanderAbramovPav/scribaWrote 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/instructions/alexanderabramovpav/scriba/agents-md)<a href="https://agentmods.dev/instructions/alexanderabramovpav/scriba/agents-md"><img src="https://agentmods.dev/badge/instructions/alexanderabramovpav/scriba/agents-md.svg" alt="Measured on agentmods" 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 | $0.02733 | $0.02733 |
| Opus 5 | $0.01367 | $0.01367 |
| Sonnet 5 | $0.00547 | $0.00547 |
| Haiku 4.5 | $0.00273 | $0.00273 |
Grade E, and why
scriba AGENTS.md scanned grade E with 4 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
> 5. Paste it here in the chat — I'll save it locally with `chmod 600` and never ask again. Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
Verify `bash`, Python 3.10+, `uv`, and `ffmpeg`/`ffprobe` are on `$PATH`. If any are missing, name the missing tool and give the exact install command for their OS (`brew install ffmpeg` on macOS, `curl -LsSf https://ast Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
- **Continue.dev** — register `bash <scriba>/scripts/transcribe.sh` as a custom slash command in `~/.continue/config.yaml`, then point the agent at this file for context. Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
Verify `bash`, Python 3.10+, `uv`, and `ffmpeg`/`ffprobe` are on `$PATH`. If any are missing, name the missing tool and give the exact install command for their OS (`brew install ffmpeg` on macOS, `curl -LsSf https://ast How it starts
The opening of the file, as written. The whole thing — 122 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — instructions for AI assistants driving scriba
This file is the tool-agnostic counterpart to SKILL.md (which is Claude Code–specific). If you're an AI agent running inside Codex CLI, Cursor, Continue, Aider, Goose, JetBrains AI, or any other coding/chat tool, and the user has asked you to help with a meeting transcription — read this.
What scriba is
scriba is a local, MIT-licensed tool that turns any audio/video file into a speaker-diarized Markdown transcript with embedded audio clips of each speaker. The CLI entry point is:
bash <scriba>/scripts/transcribe.sh <media-file>
Where <scriba> is wherever this repo was cloned. Typical paths:
~/.claude/skills/scriba/(Claude Code convention — still works for any tool)~/.config/scriba/or~/dev/scriba/(standalone install)- A relative path inside a project that vendors it
No AI is required to run scriba — the pipeline is just bash + Python. Your role as an agent is to make the experience pleasant: handle the first-run setup, surface live status without spamming, and help the user name speakers when it's done.
First-run setup — handle it for the user
Before invoking transcribe.sh the first time on a machine, check whether ~/.config/scriba/hf_token exists. If it does, skip this. If it doesn't, do not silently fall through to diarization-less mode — the user will be confused why everyone is Speaker 1.
Walk them through the HuggingFace onboarding conversationally, in their language (English by default, match the user's chat language otherwise):
To identify who's speaking, scriba needs a free HuggingFace token (~30 seconds):
- Open https://huggingface.co/join — create a free account if you don't have one.
- Open https://huggingface.co/pyannote/segmentation-3.0 and click "Agree and access repository" (one click).
- Same on https://huggingface.co/pyannote/speaker-diarization-community-1 — the model that identifies speakers (one click).
- Open https://hf.co/settings/tokens → "+ Create new token" → name it
scriba, type Read, copy the generated token (looks likehf_...).- Paste it here in the chat — I'll save it locally with
chmod 600and never ask again.
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.
- 4d ago First seen · 122 lines · 2,733 tokens per session scan E b851481e6329
scriba AGENTS.md is an instructions file published in the GitHub repository AlexanderAbramovPav/scriba (13 stars, last pushed 1mo ago), licensed MIT. It adds 2,733 tokens to every session, about $0.0137 per session on Opus 5. A static security scan graded it E with 4 findings (asks for root, downloads and executes remote code, reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
speech-swift AGENTS.md
AGENTS.md instructions for soniqo/speech-swift, covering agent instructions, communication style, workflow, running tests and benchmarks — sequential and memory-aware and git conventions.
kesha-voice-kit AGENTS.md
AGENTS.md instructions for drakulavich/kesha-voice-kit: Project rules, build commands, release process, and lessons learned live in CLAUDE.md — single source of truth for every coding agent (Claude Code, Cursor, Aider, Codex, etc.). When CLAUDE.md and any other doc disagree, CLAUDE.md wins.
delegate-local CLAUDE.md
Instructions for IsmaelMartinez/delegate-local, covering claude.md, what this repo is, commands, architecture and conventions.
kesha-voice-kit CLAUDE.md
Claude Code instructions for drakulavich/kesha-voice-kit, covering claude.md, project overview, critical development rules, default tts voices must be male and never auto-download the engine or models.
speech-swift copilot-instructions.md
Copilot instructions for soniqo/speech-swift, a project described as: AI speech toolkit for Apple Silicon — ASR, TTS, speech-to-speech, VAD, and diarization powered by MLX and CoreML.
speech-swift CLAUDE.md
Claude Code instructions for soniqo/speech-swift, a project described as: AI speech toolkit for Apple Silicon — ASR, TTS, speech-to-speech, VAD, and diarization powered by MLX and CoreML.