AReaL is an infrastructure system for training large language models with reinforcement learning, connecting model training to applications built around AI agents. Researchers and developers use it to train reasoning and agentic models through asynchronous workflows, and the catalogue add-ons support working with AReaL.
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 skills/areal-project/areal/create-prnpx skills add areal-project/AReaL --skill create-prgit clone --depth 1 https://github.com/areal-project/AReaLWrote 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/skills/areal-project/areal/create-pr)<a href="https://agentmods.dev/skills/areal-project/areal/create-pr"><img src="https://agentmods.dev/badge/skills/areal-project/areal/create-pr.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 | $0.00034 | $0.00545 |
| Opus 5 | $0.00017 | $0.00272 |
| Sonnet 5 | $0.00007 | $0.00109 |
| Haiku 4.5 | $0.00003 | $0.00055 |
Grade A, and why
create-pr 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 5d 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.
How it starts
The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Create Pull Request
Use this skill when the user asks to create or update a PR for the current branch.
Inputs
- Optional
--draft - Optional
--base <branch>(default:main)
Preconditions
- Verify the current branch is not
mainormaster. - Check for uncommitted changes with
git status --short. - Ensure
ghis available. - If there are uncommitted changes, stop and ask the user whether to commit or stash them first.
Workflow
Step 1: Check for an existing PR
- Run
gh pr view --json number,title,url,state,isDraft. - If a PR already exists, tell the user before rewriting history or force-pushing.
Step 2: Fetch and rebase
git fetch origin <base>
git rebase origin/<base>
- If rebase conflicts occur, abort the rebase and stop.
- Tell the user which files conflicted and ask them to resolve manually.
Step 3: Squash into one commit
git reset --soft origin/<base>
- Load the
commit-conventionsskill before generating the commit message. - Infer
typeandscopefrom the staged diff. - Keep the commit subject imperative and under about 72 characters.
Step 4: Generate PR title and body
- Use the squashed commit message style for the PR title.
- Follow the repository PR template at
.github/PULL_REQUEST_TEMPLATE.md. - Summarize user-facing changes, risk areas, test commands run, and skipped suites with reasons.
Step 5: Push and create or update the PR
- Push the branch.
- If history was rewritten, confirm before force-pushing.
- Create or update the PR with
gh pr createorgh pr edit. - Respect
--draftwhen requested.
Guardrails
- Never create a PR from
mainormaster. - Never silently force-push over an existing PR branch.
- Never bypass
commit-conventionsfor the squashed commit. - If
ghauthentication or remote permissions fail, stop and report the exact blocker.
Output
Report:
- Base branch used
- Final commit message
- PR title
- PR URL, if creation or update succeeded
- Any steps that were skipped or require user follow-up
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.
- 5d ago First seen · 80 lines · 34 tokens per session scan A 48f10dadf2d8
create-pr is a skill published in the GitHub repository areal-project/AReaL (5,717 stars, last pushed yesterday), licensed Apache-2.0. It adds 34 tokens to every session and 545 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
add-cuda-kernel
Rules and steps for creating a compute kernel in RL.cu project.
everything-claude-code
Development conventions and patterns for everything-claude-code. JavaScript project with conventional commits.
release
Prepare and publish stable Agent Lightning releases through the repository's version bump, pull-request checks, merge, tag, PyPI trusted-publishing, and versioned-documentation workflows. Use when asked to plan, cut, verify, or explain a release; treat nightly TestPyPI builds as a separate path.
contextual-commit
Write contextual commits that capture intent, decisions, and constraints alongside code changes. Use when committing code, finishing a task, or when the user asks to commit. Extends Conventional Commits with structured action lines in the commit body that preserve WHY code was written, not just WHAT changed.
agent-lightning
Provides the action space, tradeoffs, and evaluation context for improving an editable AI agent against a benchmark while preserving its deployment contract. Use when optimizing agent accuracy, cost, latency, or reliability.
commit
Commit current changes with a clear, descriptive message.