stenoai AGENTS.md

stenoai AGENTS.md is an instructions file for Codex, OpenCode from stenolabs/stenoai. It costs 949 tokens per session, scanned A, original, MIT.

Repository instructions for coding agents working on the stenolabs/stenoai project. They include code-review rules and guidance for bundled binaries, Python, and platform compatibility.

In plain words
What is it for?
Use them when reviewing code, calling bundled model functionality, handling Python exits, or checking changes across supported platforms.
Why use it?
They reduce the risk of changes that work during development but fail in the shipped application or break cross-platform behavior.

Instructions file for CodexOpenCode

Install

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.

agentmods
npx agentmods add instructions/stenolabs/stenoai/agents-md
Clone the repo
git clone --depth 1 https://github.com/stenolabs/stenoai

Made for: Codex, OpenCode.

Wrote 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.

agentmods badge for stenoai AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/stenolabs/stenoai/agents-md.svg)](https://agentmods.dev/instructions/stenolabs/stenoai/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/stenolabs/stenoai/agents-md"><img src="https://agentmods.dev/badge/instructions/stenolabs/stenoai/agents-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 949 This file is loaded in full into every session.
When invoked 949 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00949 $0.00949
Opus 5 $0.00475 $0.00475
Sonnet 5 $0.00190 $0.00190
Haiku 4.5 $0.00095 $0.00095

Measured 4d ago against content hash 315864895c9a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

stenoai AGENTS.md 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 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.

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.

AGENTS.md · 65 lines

How it starts

The opening of the file, as written. The whole thing — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Instructions for coding agents working in this repository.

CLAUDE.md at the repo root is the full guide: architecture, development commands, the e2e tiers, the release process, and the cross-platform rules. Read it first and treat it as authoritative. This file deliberately does not repeat it, so the two cannot drift.

The section below is different in kind. Codex's GitHub code review reads ## Code Review Rules and applies it to pull requests, so it is written for a reviewer rather than for an implementer.

Code Review Rules

These are the failure modes that are non-obvious and expensive here. Each one states the invariant and the safe path. Report a finding only when the diff actually violates the invariant, and say which input or state produces the wrong result. If a change is clean, say nothing.

Bundled binaries with an API

Never invoke a bundled binary through subprocess/spawn for an operation that also has an HTTP or library API. macOS SIP and the Electron hardened runtime strip DYLD_LIBRARY_PATH from child processes, so such a call works in development and fails only in the signed, shipped app. Model operations go through the ollama Python package. This rule is narrow: spawning the bundled stenoai CLI is the app's architecture, and ffmpeg and ollama serve have no library alternative, so none of those are findings.

Bare exit() in Python

Use sys.exit(), never bare exit(). exit is injected by site and is absent from a PyInstaller bundle, so the call raises only in the shipped app and never in development. Ruff's default rule set does not catch this.

Platform parity

The app ships on macOS and Windows from shared code, and macOS is the signed, stable build that a Windows fix must never disturb. Platform-specific behaviour must be gated on process.platform (JS) or sys.platform (Python), and platform-specific electron-builder options belong in the mac or win block rather than at the top level. The same applies to stenoai.spec: an addition that is platform-specific must be conditional, while a genuinely cross-platform one is correct unconditional. Flag an ungated change even when it is correct for the platform the author tested on.

Read the full file on GitHub · 65 lines

Changes

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.

  1. 4d ago First seen · 65 lines · 949 tokens per session scan A 315864895c9a

Subscribe to this mod's changes

stenoai AGENTS.md is an instructions file published in the GitHub repository stenolabs/stenoai (1,301 stars, last pushed today), licensed MIT. It adds 949 tokens to every session, about $0.0047 per session on Opus 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 2026-08-30.

Related

Other instructions, from other repositories