Agent Operations
How unfamiliar coding and research agents orient, select bounded work, validate it, and hand it off.
Browse topics and the exact public objects used to compile them.
How unfamiliar coding and research agents orient, select bounded work, validate it, and hand it off.
Why public eligibility must precede epistemic evaluation and how noninterference is tested.
The public project, knowledge network, reference implementation, and current operating state.
The event, provenance, federation, evaluation, disclosure, and projection architecture beneath Epistemedia.
How local-first realms exchange immutable bundles without inheriting remote verdicts or one global ontology.
Executable constitutional governance without routine discretionary human decisions or self-authorization.
How sources, spans, observations, propositions, assertions, evidence, derivations, evaluations, and projections are represented.
The human site, Markdown twins, llms.txt, API, MCP, CLI, and cross-interface identity contract.
Accepted commits, deterministic outputs, manifests, checksums, validation receipts, and deployment identity.
Falsifiable questions, benchmarks, experiments, and unresolved design tradeoffs for collective epistemic systems.
Threats including source prompt injection, evaluator collusion, Sybils, ontology poisoning, lineage laundering, and privilege escalation.