PolicyKit primary sources — 2026-05-09
What I reviewed
- Official site / docs:
- Official repository / related project docs:
- Research paper surfaced from official docs:
Key takeaways from primary sources
- PolicyKit is best understood as a governance-authoring and governance-execution layer for online communities rather than as an onchain governor or a single voting app.
- Its core abstraction is that governance can be written as policies over actions, proposals, roles, and community objects, then executed back on the platforms where a community already lives.
- The strongest reusable mechanism is the split between governable actions, trigger actions, and platform integrations. That split determines which outside behaviors a community can observe, intercept, vote on, or automatically execute.
- PolicyKit’s community model matters because one community can span multiple platforms plus a dedicated constitutional layer, which means authority can be coordinated across chat, forums, contribution tools, and policy metadata instead of remaining trapped inside one product.
- The integrations docs make clear that practical control depends on platform-specific authentication, webhook access, OAuth scopes, and revert/re-execute capability. In other words, a community’s real governance surface is partly defined by what each external platform API allows PolicyKit to see or override.
- The docs and README repeatedly position PolicyKit alongside Metagov, with PolicyKit acting as the governance-authoring application while Metagov supplies cross-platform gateway infrastructure.
Specific source signals worth retaining
- Official landing page: frames PolicyKit as an open-source tool for evolving governance in online communities and explicitly cites Elinor Ostrom as design inspiration.
- Docs landing page: says PolicyKit lets members concisely author governance procedures and automatically carry them out on home platforms like Slack, Reddit, and Discord.
- Design overview: defines the core data model around Community, CommunityPlatforms, Policy, Proposal, Role, and Action types, which is the cleanest expression of where policy logic sits.
- Integrations docs: distinguish Actions, Trigger Actions, Governable Actions, and Voting, showing how much governance power depends on external platform connectors and admin-granted credentials.
- README: describes PolicyKit as a governance-authoring application built with PolicyKit that leverages Metagov Gateway, which clarifies lineage and the driver-versus-gateway split.
- Official docs link to the ACM UIST 2020 paper, which is useful as a canonical design reference even though the active operational truth now lives more in docs and code.
Why this entry belongs in the corpus
- PolicyKit belongs in the corpus because it exposes a governance control surface that standard token-voting analyses often miss: policy authoring and execution across the actual tools a community uses day to day.
- It improves comparisons among Metagov, DAO metadata standards, chat/forum governance layers, and DAO operating systems by showing where authority migrates when governance is expressed as programmable policy over external platforms.
- The reusable insight is that online-community governance is often constrained less by the vote rule itself than by the action grammar, integration layer, and execution hooks available to the policy engine.