Adding agent files for oneDPL - #2668
danhoeflinger wants to merge 11 commits into
Conversation
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
There was a problem hiding this comment.
Pull request overview
Adds repository “agent guidance” documentation intended to help AI assistants navigate and work in the oneDPL codebase more effectively, with a concise entrypoint (AGENTS.md) and a more detailed architecture reference (IMPLEMENTATION_DETAILS.md), plus provider-specific entry files.
Changes:
- Added
AGENTS.mdwith build/test, architecture, and contribution pointers for AI agents. - Added
IMPLEMENTATION_DETAILS.mddescribing oneDPL’s layering, backend structure, and test layout. - Added
CLAUDE.mdandGEMINI.mdas thin entry files referencingAGENTS.md(with extra Gemini-specific guidance).
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 8 comments.
| File | Description |
|---|---|
| AGENTS.md | New agent-facing quick reference for building/testing, architecture, and repo conventions |
| IMPLEMENTATION_DETAILS.md | New detailed architecture/development-pattern overview for deeper context |
| CLAUDE.md | Provider-specific entry file referencing AGENTS.md |
| GEMINI.md | Provider-specific entry file referencing AGENTS.md, with additional workflow guidance |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
There was a problem hiding this comment.
Pull request overview
Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| - **Header-only library** - no binary artifacts | ||
| - **C++17 minimum** required | ||
| - **Backend selection is compile-time only** - zero runtime overhead | ||
| - **Two-phase headers** - `*_defs.h` (declarations), `*_impl.h` (implementations) |
There was a problem hiding this comment.
Do we want a requirement specifying that all AI-authored commits should be disclosed in the commit message? I think if we list it here, the agent should respect it.
It brings up another question about if we should adjust the contribution guide to give user guidance for AI generated code (e.g. all commit should be human reviewed prior to opening a PR, but I think that can be addressed separately).
There was a problem hiding this comment.
Its a good question. I think ultimately the committing user has to be responsible for the code whether or not it was partially or fully generated by an agent. In this case, I'm not so sure that adding something to the commit message is important, but its worth discussing and perhaps making a policy.
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
Co-authored-by: Matthew Michel <matthew.michel@intel.com>
Signed-off-by: Dan Hoeflinger <dan.hoeflinger@intel.com>
I have been using these agent files for some time now locally.
The goal is to provide AI agents better context with which to navigate the repository intelligently.
We use
AGENTS.mdso that it is generic across all major providers, and symlinks (@) forCLAUDE.mdandGEMINI.md. Gemini gets some extra guidance to resolve some recuring issues.We keep
AGENTS.mdsmall so that we don't bloat or confuse the context too much.We also provide and link to a more verbose description of the code layout in
IMPLEMENTATION_DETAILS.mdso it is only read into the context when needed.I'm open to other configurations, perhaps refactoring
IMPLEMENTATION_DETAILS.mdinto our main docs as a section and linking to it there, but I'd like to open it up for input from others.