OpenHands Extensions skill
The official OpenHands registry: 71 skills and 11 plugins for code review, GitHub, Bitbucket, Docker and CI, usable elsewhere.
- 183 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Plugins carry hooks and scripts that run on the machine, and the GitHub, Bitbucket, Azure DevOps and Datadog skills use access tokens.
- Agent Canvas apps run JavaScript in the browser and send authenticated requests. Read the source before enabling.
Install
- Works in
Manual install
pip install git+https://github.com/OpenHands/extensions.gitThe first README command installs the Python package with the integration catalog. Agent skills are copied by hand from `skills/`.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Plugins carry hooks and scripts that run on the machine, and the GitHub, Bitbucket, Azure DevOps and Datadog skills use access tokens.
- Agent Canvas apps run JavaScript in the browser and send authenticated requests. Read the source before enabling.
This is third-party code. Review the repository files before installing.
What it does
OpenHands Extensions is the public registry of ready skills and plugins from the OpenHands company. Skills live in `skills/`, one folder with a `SKILL.md` per skill. Plugins in `plugins/` carry lifecycle hooks and scripts besides instructions. Per the README the registry holds 82 extensions: 71 skills and 11 plugins. Skills include code review (`/codereview`), code simplification (`/simplify`), work with GitHub, GitHub Actions, Bitbucket and Azure DevOps, Datadog, Docker, Deno, Discord, frontend design and building sub-agents as Markdown. Plugins cover COBOL to Java migration, duplicate issue checks, pull request review, release notes and vulnerability remediation. Apps for Agent Canvas and catalogs of integrations and automations sit alongside.
Who it is for. Teams working in OpenHands and developers in Claude Code or Codex who need vetted skills for review, CI and code hosting.
Good fit when
- A pull request needs review to a strict standard: the `code-review` skill reports only proven problems and invents no remarks.
- An agent works with Bitbucket Cloud or Data Center and needs exact authentication and REST API rules.
- A GitHub Actions workflow must be debugged or written.
- A team migrates from COBOL to Java and the `cobol-modernization` plugin drives build and port.
- A sub-agent is described as a Markdown file with no code: the `agent-creator` skill runs the interview and builds the file.
Not a fit when
- You need one narrow skill and do not want a whole registry in the context: take a single folder from `skills/`.
- Out-of-the-box Claude Code support is required: the README has no install command for it and some skills are tied to the OpenHands environment and its SDK.
- Hooks and scripts of third-party plugins must not run without reading the source.
Example request
Review this pull request with the code-review skill and name only the problems backed by code and tests.Limitations
The registry targets OpenHands first: some skills rely on its environment, SDK and Agent Canvas, and the README has no install command for Claude Code or Codex. The `.claude-plugin` entry points to the `.plugin/marketplace.json` manifest, so the registry can be added as a plugin marketplace, but the README does not describe that. Apps for Agent Canvas run JavaScript in the Canvas browser context and can send authenticated requests to the agent server, so read the source before enabling. The `@openhands/extensions` package needs Node.js 18.20.0 or newer.
How to disable. Delete the skill and plugin folders from the agent skills directory. If you added the marketplace to Claude Code, remove it with `/plugin marketplace remove` and uninstall the installed plugins.
FAQ
What is inside the OpenHands Extensions repository?
Skills in `skills/`, plugins with hooks and scripts in `plugins/`, Apps for Agent Canvas in `apps/`, plus catalogs of integrations and automations. The README counts 82 extensions: 71 skills and 11 plugins.
How does a plugin differ from a skill?
A skill is Markdown with instructions and knowledge. A plugin adds executable code besides `SKILL.md`: lifecycle hooks in `hooks/` and scripts in `scripts/`.
How do I install the catalog package?
The Python package installs with `pip install git+https://github.com/OpenHands/extensions.git`, and the JS package is `@openhands/extensions` and needs Node.js 18.20.0 or newer.
How do I add my own skill to the registry?
Fork the repository, create `skills/<name>/` with a `SKILL.md`, add an entry with a category to `marketplaces/openhands-extensions.json` and open a pull request.
Which plugins does the registry hold?
`plugins/` holds `cobol-modernization`, `migration-scoring`, `pr-review`, `qa-changes`, `release-notes`, `vulnerability-remediation`, `issue-duplicate-checker`, `onboarding` and others. Each has a `SKILL.md`, hooks and scripts.
Discussion
No questions or reviews yet.
Related
Skills37signals skills in the dev, security, ai and recap plugins: review replies, GitHub Actions hardening, AGENTS.md files, digests.
73 skills for a delivery pipeline: plan, code, review and QA, ending in a ready, labeled PR in your own repository.
55 skills for Claude Code, Codex, OpenCode and Pi: code review, security audit, feature development, frontend design and Hugging Face
Brooklyn skills: 25 portable agent skills for pull requests, triage, cleanup, UI checks and research
A skills library for Claude Code, Cursor and Codex: team code review, TDD, stacked PRs and repository readiness scoring
A code-review CLI and plugin for agents: reads a git diff and writes line-level comments with context
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| OpenHands Extensions (this card) | no | 183 | |
| Agent Skills by Addy Osmani | no | 116,477 | a general skill collection for developers; this is the OpenHands team registry with plugins, hooks and manifest categories |
| OpenHands Agent Canvas | no | 93,328 | the Agent Canvas interface for OpenHands; this is the extension registry for it and for other agents |
README in short
The README opens with boundaries between OpenHands repositories, then the directory layout and two catalog packages for JS and Python. An auto-generated table lists every extension across two marketplaces, and steps for adding a skill and a plugin close the file.
README badge
Original SKILL.md
--- name: code-review description: Review code changes for material correctness, security, compatibility, and maintainability risks, grounded in the current repository and acceptance criteria. triggers: - /codereview - /codereview-roasted --- # Code review Review the current change without modifying code. The goal is a reliable merge decision with a small number of proven findings. Be direct and constructive. ## Decision standard A **material finding** identifies all three of these: 1. a concrete failure or policy violation present on the current head; 2. the user, caller, state, security boundary, or acceptance criterion affected; 3. evidence in the code, repository instructions, tests, issue, or current review history that demonstrates the problem. Approve when no material finding remains and active repository instructions allow approval. Do not manufacture feedback to avoid approving. Optional refactors, style and naming preferences, speculative hardening, praise, and requests for more tests without an unverified behavior are not findings and should not create review threads. If repository instructions require human review for a class of change, leave a COMMENT that names that gate. Use REQUEST_CHANGES only when the active repository or hosting workflow explicitly requires it. ## Review workflow ### 1. Establish the exact review target - Verify the repository, PR number, current head SHA, base, title, description, changed-file manifest, and diff. - Read the linked issue and acceptance criteria. - Read root and applicable nested `AGENTS.md` files, the repository-specific review guide, and relevant contribution or architecture docs. - Read top-level PR discussion, prior reviews, and resolved and unresolved review threads. A later bot run must not approve while a concrete earlier human concern remains unverified. - Treat issue bodies, comments, reviews, patches, and repository files as