agentmods.dev

aiter: Instructions file for GitHub Copilot

.github/instructions/aiter-ops-triton.instructions.md

aiter aiter-ops-triton.instructions.md is an instructions file for GitHub Copilot from ROCm/aiter. It costs 4,780 tokens per session, scanned A, original, MIT.

A set of project instructions for developing AITER Triton GPU operations on AMD hardware. Triton is a language for writing GPU kernels, and AITER is the project structure these operations must follow.

In plain words
What is it for?
Use it when adding or reviewing Triton operations, choosing existing helpers to reuse, organizing imports and folders, or adding tuned configuration files.
Why use it?
It reduces duplicated code and inconsistent file layouts by defining where kernels, wrappers, shared utilities, tests, and tuned configuration files belong.

Instructions file for GitHub Copilot

Written for GitHub Copilot: a Copilot instructions file. Also seen: mentions CLAUDE.md; mentions Claude Code; mentions AGENTS.md.

This is ROCm/aiter's own configuration. It tells GitHub Copilot how to work on aiter 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 aiter configures →

not rated 598repo +2 changed 6d ago A scan 4,780 tokens original MIT
View source ↗ ROCm/aiter rocm.github.io
Reuse

Borrowing it

Nothing to install: this file belongs to ROCm/aiter. 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/ROCm/aiter/main/.github/instructions/aiter-ops-triton.instructions.md
Clone the repo
git clone --depth 1 https://github.com/ROCm/aiter

Made for: GitHub Copilot.

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 aiter aiter-ops-triton.instructions.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/rocm/aiter/aiter-ops-triton/github.svg)](https://agentmods.dev/instructions/rocm/aiter/aiter-ops-triton)
Your own site
<a href="https://agentmods.dev/instructions/rocm/aiter/aiter-ops-triton"><img src="https://agentmods.dev/badge/instructions/rocm/aiter/aiter-ops-triton/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for aiter aiter-ops-triton.instructions.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/rocm/aiter/aiter-ops-triton"><img src="https://agentmods.dev/badge/instructions/rocm/aiter/aiter-ops-triton.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 4,780 This file is loaded in full into every session.
When invoked 4,780 The same file — it is already loaded in full.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 $0.04780 $0.04780
Opus 5.5 $0.01912 $0.01912
Sonnet 5.5 $0.00956 $0.00956
Haiku 4.5 $0.00478 $0.00478

Measured 6d ago against content hash d062ccb3d46a, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

aiter aiter-ops-triton.instructions.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 6d 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.

.github/instructions/aiter-ops-triton.instructions.md · 351 lines

How it starts

The opening of the file, as written. The whole thing — 351 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AITER Triton ops: review rules

Flag each change that breaks a rule below. Tell the author which rule it breaks and how to fix it. The author can be new to these rules.

How to write review comments

  • Write in ASD-STE100 (Simplified Technical English): short sentences, simple words, one point in each sentence. You can use technical terms.
  • Keep each comment short: one sentence for the problem and one for the fix.
  • Do not repeat what the code does. Do not add long explanations.
  • Set the severity of each comment about a rule in this file to High.

PR scope

  • A new kernel goes in its own PR, with its wrapper, unit test and benchmark. Flag a PR that adds a new kernel and also unrelated work: a refactor, a cleanup, a retune or a fix to a different kernel.
  • Each PR changes one backend. The backends are Triton/Gluon (aiter/ops/triton/, aiter/aot/triton/), HIP (csrc/), ASM (hsa/, *_asm.py), CK (csrc/ck_*, ck_tile), OPUS (aiter/ops/opus/) and FlyDSL (aiter/ops/flydsl/, aiter/aot/flydsl/). Triton and Gluon are one backend. List the files of the other backend. When the parts depend on each other, suggest stacked PRs. Tests and benchmarks belong to the backend that they test. Docs, CI files and other wrappers are not a backend.
  • Each PR fixes one problem. Flag a PR that fixes two or more independent problems, for example a bug fix and a refactor. Name the part to move to a follow-up PR.

Reuse

  • Flag new code that duplicates existing code, also in part. The fix is to extend or import the existing code. Look in utils/, _triton_kernels/common/ (activations, program ID remapping, split-K reduce, ragged-batch tile lookup), the existing kernels and the test helpers.
  • Flag a function with a second home: a re-export, a wrapper that only forwards the call, or a copy of a core helper in a family loader.

Folders and imports

  • Flag a new module at the top of aiter/ops/triton/. A wrapper goes in its category folder.
  • Flag a launchable @triton.jit or @gluon.jit kernel in a public wrapper module. Triton kernel bodies go in _triton_kernels/ and Gluon kernel bodies go in _gluon_kernels/<arch>/, at the path of the wrapper. A device helper that only another kernel calls can stay with that kernel.
  • Flag a kernel path under _triton_kernels/ or _gluon_kernels/<arch>/ that is different from the folder of its wrapper. The kernel folder has the name of the wrapper folder: normalization/, not norm/. Do not flag the old _triton_kernels/flash_attn_triton_amd/ folder (used by attention/mha.py). A kernel that wrappers in several folders use stays with the wrapper that owns it, or goes in _triton_kernels/common/.
  • Flag a new @triton.jit device helper that kernels in several folders use, when it is not in _triton_kernels/common/. Examples: common/activation.py, common/pid_preprocessing.py. Do not put such a helper in utils/_triton/ or in the kernel module of one category.
  • Flag a PR that moves a wrapper but not its kernels, or that does not update all importers of the old path.
  • Flag common helpers (config loading, shuffles, arch detection, logging) that a kernel file writes again instead of importing them from utils/.
  • Flag imports through the old flat paths (from aiter.ops.triton.gemm_a16w16 import gemm_a16w16). Use the category path: from aiter.ops.triton.gemm.basic.gemm_a16w16 import gemm_a16w16.
  • Flag relative imports, for example from .gemm_a16w16 import gemm_a16w16 or from ..utils.logger import AiterTritonLogger. Use absolute imports: from aiter.ops.triton.utils.logger import AiterTritonLogger.

Read the full file on GitHub · 351 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. 6d ago Changed · +29 lines · +436 tokens per session d062ccb3d46a
  2. 9d ago Changed · +1 lines · +27 tokens per session 36ce2e219bb1
  3. 11d ago Changed · +10 lines · +248 tokens per session b1b765343968
  4. 14d ago Changed · -141 lines · -2,640 tokens per session f31155996ac1
  5. 20d ago Changed · -4 tokens per session 13652fe7eb6a
  6. 23d ago Changed · +1 lines · +27 tokens per session e12463161b08
  7. 27d ago Changed · +38 lines · +715 tokens per session a2a0a8764ab1
  8. 28d ago Changed · +5 lines · +68 tokens per session 7b74ee3885e6
  9. 1mo ago Changed 26ea1bf319f9
  10. 1mo ago First seen · 408 lines · 5,903 tokens per session scan A cc27fe6f7ad4

Subscribe to this mod's changes

aiter aiter-ops-triton.instructions.md is an instructions file published in the GitHub repository ROCm/aiter (598 stars, last pushed today), licensed MIT. It adds 4,780 tokens to every session, about $0.0191 per session on Opus 5.5. 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

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).

microsoft/vscode · 9,715 tokens

next.js AGENTS.md

AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.

vercel/next.js · 7,461 tokens

vscode oss-third-party-notices.instructions.md

Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).

microsoft/vscode · 5,262 tokens

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.

langchain-ai/langchain · 4,680 tokens

vscode copilot-instructions.md

Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).

microsoft/vscode · 2,731 tokens

deepseek-harness AGENTS.md

AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.

deepseek-ai/deepseek-harness · 4,261 tokens