granite.build: Skill for Claude Code

.claude/skills/create-build/SKILL.md

create-build is a skill for Claude Code from ibm-granite/granite.build. It costs 106 tokens per session (4,911 once invoked), scanned A, original, Apache-2.0.

A guide for creating a Granite.build YAML file that runs a compute job, such as model training, inference, data processing, evaluation, or another script.

In plain words
What is it for?
It helps define inputs, outputs, resources, settings, and the command that runs a workload on Granite.build.
Why use it?
It gives these jobs one self-contained run definition instead of requiring a separate step directory and scattered configuration.

Skill for Claude Code

Written for Claude Code: allowed-tools in frontmatter.

This is ibm-granite/granite.build's own configuration. It tells Claude Code how to work on granite.build itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything granite.build configures →

Reuse

Borrowing it

Nothing to install: this file belongs to ibm-granite/granite.build. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/ibm-granite/granite.build/main/.claude/skills/create-build/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/ibm-granite/granite.build

Made for: Claude Code.

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 create-build

README.md
[![agentmods](https://agentmods.dev/badge/skills/ibm-granite/granite.build/create-build.svg)](https://agentmods.dev/skills/ibm-granite/granite.build/create-build)
Your own site
<a href="https://agentmods.dev/skills/ibm-granite/granite.build/create-build"><img src="https://agentmods.dev/badge/skills/ibm-granite/granite.build/create-build.svg" alt="Measured on agentmods" height="20"></a>
Per session 106 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,911 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
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 5.1 $0.00106 $0.04911
Opus 5 $0.00053 $0.02455
Sonnet 5 $0.00021 $0.00982
Haiku 4.5 $0.00011 $0.00491

Measured 3d ago against content hash 6075b21789be, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

create-build scanned grade A with 1 finding 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 3d 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

allowed-tools: Bash(ls *) Bash(test *) Bash(cat *) Bash(grep *) Bash(find *) Bash(curl *) mcp__gbmcp__gbserver_status mcp__gbmcp__gbserver_start mcp__gbmcp__build_start mcp__gbmcp__build_status mcp__gbmcp__build_log mcp_
.claude/skills/create-build/SKILL.md · 235 lines

How it starts

The opening of the file, as written. The whole thing — 235 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Author a Granite.build build

This skill makes a workload (train a model, run inference, generate/process data, evaluate, or any script) run on Granite.build. In the standalone bash environment you produce one artifact:

  • A build.yaml — the run definition. It references the built-in command step by URI, wires inputs/outputs, passes per-run settings, and carries the workload itself inline in the command step's config.command_config.command. For anything more than a one-liner, the command writes its own script via a shell heredoc and runs it. Keep the build.yaml with the user's project (e.g. a lora-granite/build.yaml folder).

You do not author a step directory (step.yaml + bash_scripts/) for the common case — the generic command step already exists and runs whatever command you give it. This keeps the whole workload self-contained in one file.

Granite.build is workload-agnostic: training, inference, data-gen, and evaluation are all "a command that runs." What changes between workloads is the command string and the settings in build.yaml — not the structure.

Scope — this skill covers ONLY the inline command + heredoc approach, and delegates everything more complex to create-step. The heredoc approach is bash-specific (it relies on the bash environment's runtime contract — see below) and is not portable across environments: Docker/LSF/skypilot provision Python + dependencies and wire inputs differently. For the standalone-bash default it's the simplest, fastest path. The moment you need a reusable, packaged, or owned-code step — its own step.yaml + bash_scripts/, referenced from the build by a file:// URI — stop and use the create-step skill. This skill does not describe step authoring at all; it routes you to an existing step (below) or hands the case off to create-step.

First: is there already a step for your workload? (prefer it over inline)

The standalone space ships tested, purpose-built steps. If one matches your workload, use it — it's battle-tested, and reinventing it inline is slower and more error-prone. Check these before reaching for a heredoc:

Read the full file on GitHub · 235 lines

Files

What ships with it

4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. 3d ago First seen · 235 lines · 106 tokens per session scan A 6075b21789be

Subscribe to this mod's changes

create-build is a skill published in the GitHub repository ibm-granite/granite.build (44 stars, last pushed 3d ago), licensed Apache-2.0. It adds 106 tokens to every session and 4,911 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-04.

Related

Other skills, from other repositories

tensorrt-llm

High-throughput LLM inference on NVIDIA GPUs.

NousResearch/hermes-agent · 18 tokens

google-cloud-solution-guided-gke-ai-migration

Guides the migration of existing AI workloads (Cloud Run, Gemini API, Gemini Enterprise Agent Platform) to self-hosted GKE inference using gcloud and kubectl. Use when the user has an existing AI inference workload (on Cloud Run, the Gemini API, Gemini Enterprise Agent Platform, or a custom VM) and wants to move it to…

google/skills · 157 tokens

agent-platform-tuning

Agent Platform Model Tuning. Use when you need to fine-tune open models or Gemini models using Agent Platform infrastructure. Don't use for model training outside Agent Platform, model deployment to endpoints (use agent-platform-deploy), or managing serving endpoints (use agent-platform-endpoint-management).

google/skills · 64 tokens

agent-platform-endpoint-management

Manages Agent Platform serving endpoints. Use when you need to create, list, describe, update, or delete serving endpoints for model deployment on Agent Platform. Also use when troubleshooting endpoint permission, quota, or resource busy errors. Don't use for deploying models to endpoints or for running model…

google/skills · 64 tokens

gke-inference

Deploys and optimizes AI/ML inference workloads on GKE, using GPUs, TPUs, and model servers. Use when deploying GKE inference servers, configuring GKE GPU resources for inference, or deploying LLMs on GKE. Don't use for generic batch jobs or HPC task queues (use gke-batch-hpc instead).

google/skills · 74 tokens

modal

Modal is a serverless cloud platform for running Python on demand, including on-demand GPUs. Use when deploying or serving AI/ML models, running GPU-accelerated workloads (training, fine-tuning, inference), serving web endpoints, scheduling batch jobs, or scaling Python code to cloud containers with the Modal SDK.

K-Dense-AI/scientific-agent-skills · 65 tokens