Marin is an open-source research program, software platform, and community for developing foundation models such as large language models. Researchers use it for data preparation, tokenization, pretraining, posttraining, evaluation, and related experiments, including work on audio-text, DNA, and protein models. The catalogue entries are add-ons that support workflows around Marin.
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/marin-community/marin/agents-mdgit clone --depth 1 https://github.com/marin-community/marinWrote 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/marin-community/marin/agents-md)<a href="https://agentmods.dev/instructions/marin-community/marin/agents-md"><img src="https://agentmods.dev/badge/instructions/marin-community/marin/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 | $0.02840 | $0.02840 |
| Opus 5 | $0.01420 | $0.01420 |
| Sonnet 5 | $0.00568 | $0.00568 |
| Haiku 4.5 | $0.00284 | $0.00284 |
Grade A, and why
marin 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 5d 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 — 248 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent Guidelines for Marin
Start with the shared practices below. Consult subproject manuals for directory-specific guidance:
lib/levanter/AGENTS.md— Levanter (JAX training library)lib/marin/AGENTS.md— Marin (pipeline framework)lib/iris/AGENTS.md— Iris (job orchestration)lib/zephyr/AGENTS.md— Zephyr (dataset processing)lib/fray/AGENTS.md— Fray (distributed execution)
Operational Guides
For debugging and operating live infrastructure, read the relevant OPS.md:
lib/iris/OPS.md— cluster lifecycle, job/task management, profiling, SQL queries, GCP/CoreWeave operationslib/zephyr/OPS.md— pipeline debugging, straggler diagnosis, coordinator queries, diagnostic patterns
Zephyr OPS.md references Iris OPS.md for shared infrastructure commands — read Iris first when debugging zephyr jobs on Iris.
Infrastructure (Pulumi)
infra/ hosts several independent Pulumi projects following three distinct
patterns (infrastructure, application deploys, SaaS resource declarations). Read
infra/pulumi.md before creating or modifying a Pulumi project so new work
lands in the right pattern.
Workflow Playbooks
Skills are task-focused playbooks in .agents/skills/ (also accessible as
.claude/skills/). Before starting any non-trivial task, check whether a
matching skill exists by scanning the skill descriptions in your system
prompt. If a skill matches, invoke it via the Skill tool — do not skip it in
favor of ad-hoc commands.
Handle Requests
If a request comes from Slack or GitHub and appears to be a simple question, you may answer it in the originating conversation instead of making a repository change. Otherwise, carry the request through the applicable change and landing workflow; do not stop after investigation while a safe, in-scope fix remains.
Search Prior Work
Use Echo when prior Marin decisions, incidents, workflows, GitHub work, or indexed repository documentation could inform a task:
uv run infra/echo/cli.py search "how do I deploy Iris"
uv run infra/echo/cli.py search "compare cache implementations" --repository all
uv run infra/echo/cli.py get <source-id>
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.
- 5d ago First seen · 248 lines · 2,840 tokens per session scan A f41441ef19b6
marin AGENTS.md is an instructions file published in the GitHub repository marin-community/marin (3,354 stars, last pushed today), licensed Apache-2.0. It adds 2,840 tokens to every session, about $0.0142 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
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.