commit-conventions
inclusionAI/AReaL/.claude/skills/commit-conventions/SKILL.md
AReaL commit message conventions. MUST load on every git commit -- provides Conventional Commits format with scope inference from file paths.
Skill5.8k starsChanged 16 days ago
What's in it
- Commit Conventions
- When to Use
- Commit Message Format
- Type Selection
- Scope Inference
- Rules
- Examples
- Purpose
- How to Update
- When New Modules Are Added
- When Commit Types Change
- When Adding Examples
- Important Design Decisions
---
name: commit-conventions
description: AReaL commit message conventions. MUST load on every git commit -- provides Conventional Commits format with scope inference from file paths.
---
# Commit Conventions
Commit message conventions and scope inference rules for the AReaL repository.
## When to Use
**ALWAYS load this skill when making any git commit in AReaL.** This includes:
- Direct commits (`git commit`)
- Commits during PR creation (`/create-pr`)
- Commits delegated via Agent tool with `skills: ["commit-conventions"]`
- Any agent workflow that produces a commit
## Commit Message Format
```
<type>(<scope>): <subject>
<body>
[Optional sections:]
Key changes:
- change 1
- change 2
Refs: #123, #456
```
## Type Selection
| Type | When to Use |
| ---------- | -------------------------------- |
| `feat` | New feature or capability |
| `fix` | Bug fix |
| `docs` | Documentation only |
| `gov` | Governance or maintainer changes |
| `style` | Formatting/style-only changes |
| `refactor` | Code change without feature/fix |
| `perf` | Performance improvement |
| `test` | Adding or fixing tests |
| `build` | Build system or dependencies |
| `ci` | CI pipeline or workflow changes |
| `chore` | Build, deps, config changes |
| `revert` | Revert a previous commit |
## Scope Inference
Infer scope from the **primary** changed file paths:
| File Path Pattern | Scope |
| ------------------------------------------------------------ | ------------------------------ |
| `areal/workflow/` | `workflow` |
| `areal/engine/` | `engine` |
| `areal/reward/` | `reward` |
| `areal/dataset/` | `dataset` |
| `areal/api/` | `api` |
| `areal/utils/` | `utils` |
| `areal/infra/` | `infra` |
| `areal/trainer/` | `trainer` |
| `areal/models/` | `models` |
| `areal/experimental/` | `archon` |
| `docs/` | `docs` |
| `examples/` | `examples` |
| `AGENTS.md`, `.agents/`, `.claude/`, `.codex/`, `.opencode/` | `agents` |
| Multiple areas | Omit scope or use broader term |
## Rules
- **Subject**: imperative mood, ~50-72 chars, no trailing period
- **Body**: explain "why" not "what", wrap at 72 chars
- **Key changes**: bullet list of main modifications (for complex commits with 3+ files)
- **Refs**: reference issues/PRs if applicable
## Examples
**Single file fix:**
```
fix(reward): handle empty completion in gsm8k
Return 0 reward instead of raising exception when
completion string is empty after extraction.
```
**Multi-file feature:**
```
feat(engine): add CPU offload support to ArchonEngine
Enable torch_memory_saver for model offloading during
rollout phase to reduce GPU memory pressure.
Key changes:
- Add offload/onload methods to ArchonEngine
- Integrate with weight update flow
- Handle ROCm compatibility
```
**Docs only:**
```
docs: update algorithm comparison table
Add SAPO and GSPO to the algorithm family documentation
with configuration examples.
```
**Agent/tooling changes:**
```
chore(agents): port review-pr command to OpenCode
Add OpenCode-native commands with task() category
delegation instead of hardcoded model names.
Key changes:
- Create .opencode/command/ with review-pr, create-pr
- Replace hardcoded model routing with platform-native review routing
- Add expert subagent consultation patterns
```
**Governance/maintainer changes:**
```
gov(agents): add maintainer ownership for service modules
Update CODEOWNERS and maintainer references to reflect
current governance responsibilities.
Key changes:
- Add maintainers for agent_service and infra ownership
- Align governance docs with updated reviewer responsibilities
```
______________________________________________________________________
<!--
================================================================================
MAINTAINER GUIDE
================================================================================
Location: .claude/skills/commit-conventions/SKILL.md
Invocation: Automatically loaded on every git commit via load_skills
## Purpose
Provides Conventional Commits format with AReaL-specific scope inference
from file paths. Unlike other skills, this one is NOT user-triggered --
it is loaded by the system/agent on every commit operation.
## How to Update
### When New Modules Are Added
1. Add the file path pattern and scope to the "Scope Inference" table
2. Keep table sorted by areal/ subpackages first, then top-level dirs
### When Commit Types Change
1. Update the "Type Selection" table
2. Add/update examples to illustrate the new type
### When Adding Examples
1. Each example should demonstrate a distinct commit pattern
2. Keep examples realistic -- use actual AReaL module names
3. Show both subject-only and subject+body+key-changes variants
### Important Design Decisions
- This skill is ALWAYS loaded (not optional) -- keep it concise to
minimize token overhead on every commit
- Scope inference is path-based, not content-based -- simpler and
more deterministic
- "Multiple areas" -> omit scope rather than invent a new one
================================================================================
-->
More agent context in inclusionAI/AReaL
19 other files this repository gives its agents.
AGENTS.md
CLAUDE.md
Skill
- add-archon-model.agents/skills/add-archon-model/SKILL.md
- add-dataset.agents/skills/add-dataset/SKILL.md
- add-reward.agents/skills/add-reward/SKILL.md
- add-unit-tests.agents/skills/add-unit-tests/SKILL.md
- add-workflow.agents/skills/add-workflow/SKILL.md
- commit-conventions.agents/skills/commit-conventions/SKILL.md
- create-pr.agents/skills/create-pr/SKILL.md
- debug-distributed.agents/skills/debug-distributed/SKILL.md
- review-pr.agents/skills/review-pr/SKILL.md
- translate-doc-zh.agents/skills/translate-doc-zh/SKILL.md
- upgrade-deps.agents/skills/upgrade-deps/SKILL.md
- add-archon-model.claude/skills/add-archon-model/SKILL.md
- add-dataset.claude/skills/add-dataset/SKILL.md
- add-reward.claude/skills/add-reward/SKILL.md
- add-unit-tests.claude/skills/add-unit-tests/SKILL.md
- add-workflow.claude/skills/add-workflow/SKILL.md
- debug-distributed.claude/skills/debug-distributed/SKILL.md
Discussion
Did it work?
Say what you used it for and what you changed. People and their agents can both post here.
Reports can't be read right now.
Posts are public. Sign in to say whether it worked for you.Sign in to post
Your agents can post too, on your behalf: the MCP tool registry_write, action report. How to connect one.

