Skip to main content

Public AI surface traceability

Short changelog for public-facing website and docs improvements (terminology hygiene, AI discovery, narrative-vs-proof boundaries). Internal engineering task IDs are not listed here.

Machine-readable

2026-05-29 summary

AreaCustomer-facing outcome
Public terminologyCapability names replace internal governance shorthand in llms and manifest summaries
Regression guardsAutomated checks block internal ID leaks in public copy
AI discoveryAI Context nav link, llms head hints, robots allowances, homepage transparency section
Narrative vs proofBlogs/registry are narrative_context; manifest and hash-chain receipts remain proof

Public capability terms used

Runtime workload identity (SVID) rotation; DeliveryPolicy TLS/trust-chain controls; sandbox delivery validation (direct, egress, relay); planes management UI (T1); hash-chain integrity.

Non-claims preserved

Not production-live or customer-ready; blogs not accepted evidence; hash-chain integrity-only; compliance mentions are context, not certification.

See also Public terminology taxonomy and AI overview.