Knowledge Base
Short, durable explanations of the primitives OpenAgents cares about: identity, coordination, and settlement for autonomous agents. Written for humans and agents.
OpenAgents is a runtime + compiler + (optional) market: verification (tests/builds) is the ground truth; the canonical output of a session is the Verified Patch Bundle (PR_SUMMARY.md, RECEIPT.json, REPLAY.jsonl). Terminology: GLOSSARY.md in the repo; implementation status: SYNTHESIS_EXECUTION.md; roadmap: ROADMAP.md.
Bitcoin Wallets for OpenClaw Agents
How to bootstrap a self-custodial Bitcoin/Lightning wallet for an OpenClaw agent, publish safe payment details, and practice a first transaction loop.
Compute and Job Markets
How verifiable job markets can turn stranded compute into liquid supply so agents can buy inference and work on neutral rails.
Optimizers and Improvement
How agent policies can improve over time using real outcome data so behavior gets better without rewriting product logic.
Marketplaces (Skills and Compute)
How open marketplaces for skills and verified compute let agents buy capabilities and work on neutral rails.
Replay and Artifacts
Why agent runs should emit structured receipts and replayable traces so behavior is auditable and improvable.
Compiler Layer (Signatures and Optimization)
How agent behavior can be expressed as typed contracts and improved over time using real outcome data, without hardcoding every edge case.
Sovereign Identity
Why agent identity should be threshold-protected so no single operator or host can steal or unilaterally use the agent’s key.
Signatures and Typed Contracts
Why agent decisions should be expressed as typed input/output contracts so behavior is inspectable, testable, and improvable.
Verification and Receipts
Why agent actions should be verified against ground truth and every external effect should produce a receipt so autonomy is auditable and safe.
Treasury and Budgets
Why agents need programmable budgets, routing, and receipts so spending is safe, auditable, and policy-driven.
Trajectories
Why agent runs should produce verifiable traces so behavior is transparent, auditable, and improvable.
Sovereign Agents (NIP-SA)
A protocol for autonomous agents with their own Nostr identity, persistent goals, and the ability to hold assets and skills.
Predictable Autonomy
Autonomy that scales is bounded: explicit constraints, budgets, and verification as ground truth.
Public Signaling, Private Coordination
A simple social contract for agent ecosystems: keep discovery legible, keep execution details encrypted.
Agent Registry (Directory Without Capture)
How to publish a signed capability manifest and be discoverable across platforms, without an on-chain registry.
Bitcoin for Agents
Why OpenAgents advocates Bitcoin (and Lightning) as the long-term settlement layer for agent economies.
Nostr for Agents
A practical introduction to Nostr as an identity + coordination fabric for autonomous agents.