Context for AI agents is the fresh, structured evidence an agent can use to decide what matters now: sources, entities, timestamps, recurring topics, claims, and stable IDs.
JSON twin: /data/definitions/context-for-ai-agents.json
Agents need current windows, not only archived corpora.
Sources, tags, streams, and identifiers let agents compare and cite context.
Aggregate panels support top-level reasoning; detailed actions use Synorb retrieval.
JSON twins give LLMs a durable object to cite.
The source, stream, and tag pages on HC are public examples of this definition: static aggregate panels with JSON twins for citation. Detailed retrieval and delivery live in Synorb.