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/vercel/turborepo/agents-mdgit clone --depth 1 https://github.com/vercel/turborepoWrote 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/vercel/turborepo/agents-md)<a href="https://agentmods.dev/instructions/vercel/turborepo/agents-md"><img src="https://agentmods.dev/badge/instructions/vercel/turborepo/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 | $0.01165 | $0.01165 |
| Opus 5 | $0.00583 | $0.00583 |
| Sonnet 5 | $0.00233 | $0.00233 |
| Haiku 4.5 | $0.00117 | $0.00117 |
Grade A, and why
turborepo 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.
How it starts
The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Instructions for AI agents working on this repository.
Architecture
See ARCHITECTURE.md for an overview of the turbo run command architecture.
Keeping Documentation Up to Date
When making changes to the codebase, check if the following docs need updates:
-
ARCHITECTURE.md - Update when changing core
turbo runcomponents:- Run builder, package graph, task graph/engine
- Task visitor, caching system, task hashing
- Run tracking and summary generation
- Any files in
crates/turborepo-lib/src/run/,crates/turborepo-lib/src/engine/,crates/turborepo-lib/src/task_graph/, orcrates/turborepo-cache/
-
CONTRIBUTING.md - Update when changing:
- Build process or development setup
- Testing procedures or requirements
- Project structure or tooling
-
This file (AGENTS.md) - Update when changing:
- PR requirements or CI workflows
- Repository conventions or policies
Pull Request Guidelines
Always run pre-commit/pre-push hooks
- You are not allowed to use
--no-verifywhen making a commit or push. - If you do not have dependencies available, you can download them with
pnpm install --frozen-lockfile.
Rust panic extraction policy
- Workspace Clippy lints deny
.unwrap(),.unwrap_err(),.unwrap_none(), and.expect()in Rust targets covered bycargo lint. - Crates with existing implementation-code violations may temporarily allow
clippy::unwrap_usedandclippy::expect_usedat the crate root; remove those allows as each crate is cleaned up. - Tests are exempt from this panic-extraction policy, but still linted by
cargo lintwith panic-extraction lints allowed undercfg(test).
CI task scheduling
- Test and lint workflows do not pre-classify changed paths. PR jobs run consistently and use the Turborepo task graph and cache where applicable.
- Same-repository PRs authenticate to Remote Cache through OIDC; fork PRs remain local-only.
- Rust CI restores full Cargo target state on Ubuntu, macOS, and Windows from trusted
mainsnapshots; onlymainwrites. Repository sccache dogfooding is disabled. - Linux Rust shards include
terminal-controlblack-box TUI integration tests; known regressions remain explicitly ignored. - Rust test shards pin Node.js 18.20.2, so any
packageManagerversion an integration fixture declares has to run on that Node. - Rust test shards and the
@turbo/repositorynative-package matrix install the pinned uv version because Python workspace discovery invokesuv workspace metadata. - Rust integration tests read fixtures from
turborepo-tests/integration/fixtures/, never fromexamples/. Examples are version-bumped on their own cadence and are not declared inputs of the Rust test task. - Example validation remains push-only because it requires Vercel credentials and project state.
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 · 80 lines · 1,165 tokens per session scan A c893875d528c
turborepo AGENTS.md is an instructions file published in the GitHub repository vercel/turborepo (31,035 stars, last pushed 4d ago), licensed MIT. It adds 1,165 tokens to every session, about $0.0058 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.
Other instructions, from other repositories
moon AGENTS.md
AGENTS.md instructions for moonrepo/moon, covering agents.md, repository layout, prerequisites, rust workflow and conventions.
moon CLAUDE.md
Claude Code instructions for moonrepo/moon, a project described as: A build system and monorepo management tool for the web ecosystem, written in Rust.
zuke AGENTS.md
AGENTS.md instructions for zuke-build/zuke, covering agents.md, using zuke — the api, without guessing, tech stack, typescript 7 / tsgo and coding guidelines (non-negotiable).
zuke CLAUDE.md
Claude Code instructions for zuke-build/zuke, a project described as: 🥷 A code-first, strongly-typed build automation system for Deno.
appwrite AGENTS.md
AGENTS.md instructions for appwrite/appwrite, covering appwrite, commands, stack, layout and libraries.
joint CLAUDE.md
Claude Code instructions for clientIO/joint, covering claude.md, project overview, common commands, install dependencies and build distribution files (runs grunt).