# Global Agent Brief This brief summarizes the current canon service surface for agents. - Infospace slug: `canon` - Artifact count: 44 - Retrieval index items: 44 - Primary confidence command: `make validate` - Refresh generated indexes and views with: `make index` - Refresh agent briefs and interface templates with: `make agent-briefs` ## Useful Commands - `PYTHONPATH=src python3 -m info_tech_canon inspect` - `PYTHONPATH=src python3 -m info_tech_canon validate` - `PYTHONPATH=src python3 -m info_tech_canon graph` - `PYTHONPATH=src python3 -m info_tech_canon index` - `PYTHONPATH=src python3 -m info_tech_canon profile validate small-saas` ## Retrieval Entry Points - `agent/retrieval-index.md` - `agent/retrieval-index.yaml` - `agent/retrieval-index.json` - `agent/briefs/` for per-artifact briefs - `agent/templates/canon-interface-card.template.yaml` ## Common Distinctions - **Actor vs Subject vs Principal**: Use actor for the acting entity in a context, subject for the entity a policy evaluates, and principal for the authenticated identity bound to access decisions. - **Organization Role vs AccessRole vs CARING role**: Organization roles describe responsibility or position; access roles describe permissions; CARING roles classify access-governance needs and analysis. - **Policy vs Control vs Evidence**: Policy states intent or rule, control implements or enforces that rule, and evidence records why the claim should be trusted. - **Intent vs Scope vs Purpose**: Intent captures why a producer or consumer exists, scope bounds producer ownership and promises, and purpose captures consumer demand anchored in consumer intent. ## Consumption Notes - Treat `seeds/` as provenance. - Treat `infospace/` as the service-consumable canon root. - Generated files are marked and can be refreshed deterministically.