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/ibm-granite/granite.build/agents-mdgit clone --depth 1 https://github.com/ibm-granite/granite.buildWrote 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/ibm-granite/granite.build/agents-md)<a href="https://agentmods.dev/instructions/ibm-granite/granite.build/agents-md"><img src="https://agentmods.dev/badge/instructions/ibm-granite/granite.build/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 5.1 | $0.04280 | $0.04280 |
| Opus 5 | $0.02140 | $0.02140 |
| Sonnet 5 | $0.00856 | $0.00856 |
| Haiku 4.5 | $0.00428 | $0.00428 |
Grade A, and why
granite.build 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 today.
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 — 215 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
This file provides guidance to coding agents (Claude Code and others) when working with code in this repository.
Git Commits
Always create commits with the -s (sign-off) flag — e.g. git commit -s -m "...". The upstream repo (ibm-granite/granite.build) enforces the DCO check, which requires a Signed-off-by: trailer on every commit. Commits without it will fail CI on the PR.
GitHub PRs and Issues
When writing PR or issue descriptions, comments, and reviews, do not use bare #-number notation (#1, #2, …) to enumerate list items or steps. GitHub auto-links #N to the issue or PR with that number, so a list numbered this way turns into a set of misleading cross-references. Use plain numbers (1., 2.), letters ((a), (b)), or descriptive names instead. Reserve #N for genuine references to a specific issue or PR (e.g. "builds on #257").
Project Overview
gbserver is the build orchestration server for LLM.Build (Granite.Build). It manages model build pipelines — watching PRs and repos for build configurations, executing multi-step builds on Kubernetes/LSF clusters, and exposing a REST API for build management. Written in Python 3.11+, it uses Click for CLI, FastAPI for the REST API, and SQLAlchemy with PostgreSQL for metadata storage.
Common Commands
Virtual Environment Setup
# Requires ARTIFACTORY_USER and ARTIFACTORY_API_KEY env vars
make venv
source .venv/bin/activate
Running Tests
pytest -s test runs the suite (requires GBTEST_SPS_IBMCLOUD_API_KEY for secret retrieval); narrow with the usual test/dir, file.py, or file.py::Class::method args. The non-obvious make targets:
make cicd-pr-test # abbreviated CI set (coverage + parallel)
make cicd-merge-test # extended CI set (the `extended` marker)
# `-setup` targets provision the venv and infra first:
make quick-tests-setup quick-tests # fast: GBTEST_MODE=mock, -m "not ibm and not extended"
make extended-tests-setup extended-tests # full: GBTEST_MODE=live, -m "not ibm"; setup also brings up MinIO + SLURM
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.
- today First seen · 215 lines · 4,280 tokens per session scan A 403cbd5f86ba
granite.build AGENTS.md is an instructions file published in the GitHub repository ibm-granite/granite.build (44 stars, last pushed today), licensed Apache-2.0. It adds 4,280 tokens to every session, about $0.0214 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 2026-09-04.
Other instructions, from other repositories
daisyui code_generation_with_git.instructions.md
Instructions for saadeghi/daisyui, covering activation, initial setup, development loop, error handling and session end.
planning-with-files AGENTS.md
AGENTS.md instructions for OthmanAdi/planning-with-files, covering agents.md — planning-with-files agent reference card, commit rules, release checklist (12 steps), version bump scope and changelog format.
Commit Message
Comprehensive Conventional Commits guidelines for openwebui-extensions.
dario CLAUDE.md
Claude Code instructions for askalf/dario, covering dario — notes for claude code (and other llm coding agents), commits and prs, style, scope and when in doubt, ask.
ai-git-commiter CLAUDE.md
Instructions for zdt1013/ai-git-commiter, covering ai git commiter, 项目结构, 技术栈, 常用命令 and 安装依赖.
orkid CLAUDE.md
Claude Code instructions for tweakoz/orkid, covering orkid — repo-wide agent law and commits.