You can bring the personal agent you already use to 1Hive. Let it help you understand the community, find work that fits your interests, and contribute to the commons.
The starting point is a shared wiki, an open repository, and a small library of reusable workflows. You choose your agent and keep control of its accounts, permissions, and actions. This approach does not require a shared model account or handing model credentials to 1Hive.
Give your agent a starting point
Point your agent at AGENTS.md. It links to the community context, contribution conventions, and skills library.
You can start with a request like this:
Read https://1hive.org/AGENTS.md and help me understand 1Hive. My interests are [your interests]. Use the shape-contribution skill to investigate a meaningful contribution across infrastructure, communications, or public goods. Show me evidence, a reusable output, and a bounded first experiment before publishing anything or acting on my behalf.
If your agent cannot read websites, use the repository entry point or give it the relevant files. The skills are plain Markdown instructions, not software you must install.
Start with useful work
The First contribution launchpad helps you originate work in three lanes: core infrastructure, communications and narrative, and tangible public goods. Use the shape-contribution skill to compare possible contributions and prepare a concept with a clear beneficiary, evidence, resource limit, and first experiment.
Choose a bounded result with a purpose another person can understand:
- Learn the community. Ask for a map of the mission, Honey, governance, and places to contribute, connected to your interests.
- Build the foundation. Improve a Honey tool, agent workflow, skill, or documented process that other contributors can reuse.
- Produce a public good. Explore a bounded research question, dataset, evaluation environment, or open recipe with appropriate source rights and review.
- Follow public activity. Ask for a summary of a specific period of public discussions, issues, or proposals, with links and a clear account of what could not be checked.
- Help someone ship. Review a contribution, reproduce an issue, or explain a tool before starting something new.
The skills library provides initial workflows for orientation, wiki improvements, and public-activity summaries. They can evolve as contributors learn what works.
Keep people in control
Agree on what your agent can read, prepare, and do. Reading public information and drafting a change are different from posting a message, opening a pull request, signing a transaction, or making a commitment.
Review the sources and proposed actions. Keep private messages and credentials out of public contributions. Let others know when an agent helped with submitted work, and take responsibility for its accuracy. Your agent should follow the norms of each space and avoid duplicate issues, repetitive replies, and bulk-generated content.
Useful autonomy comes from a clear scope, reliable feedback, and work people can review. Contribute compute only with the account owner’s permission and within provider terms; never share model credentials. Agent review can help expose errors, but consequential, unresolved, or specialist questions need human judgment.
Contribute back to the shared foundation
The website repository is a place to improve the wiki, agent instructions, and skills together. Follow the contribution guide, check existing work, and submit a small change for review when you are ready and have authorized publication.
For work that needs a longer-lived plan, your agent can help prepare an improvement proposal with sources, a delivery plan, risks, and checks. The process is a working draft. A small wiki fix can go straight through the normal pull-request workflow.
Good skills explain when to use them, what inputs they need, how to verify a result, what to return, and when to stop for human input. A better workflow should help the next contributor as well as the person who created it.
Where this could lead
The longer-term direction is cooperative participation: personal agents helping people stay informed, coordinate work, learn from one another, and contribute more effectively to public goods.
Recurring summaries, opt-in notifications, contribution records, reputation, and connections to Honey rewards are possibilities to explore. This first foundation does not provide those automated services or establish a new rewards program. Any funding or reward still depends on the relevant community process and decisions. See Curating community contributions for funding references and what to verify before relying on them.
Start with Our mission and values, explore How you can contribute, or use Navigating core community infrastructure to find the right place for your work.