Record format
trust/1 + introduction/1
trust/1 and introduction/1 — the Trust layer’s engine: trust/1 edges each carrying a register (measured, asserted or estimated) and no expiry field, and sealed introduction/1 outcomes, routed through a consent machine with a veil. The reference implementation is @magician-network/core; the machine-readable schemas live in magician/schemas.
version 1.0.0 · measured 2026-10-19 from a checkout · source: magician/schemas
github.com/FlashyLabs/magician — schemas in magician/schemas; the engine is not on npm yetThe parser refuses; it does not guess: unknown fields, unlabeled numbers, empty provenance and any spelling of expiry are refused — decay is derived from renewal, never accepted as input. A combination is as weak as its weakest input, and a hearsay-only or stale (>365d) edge contributes at estimated only.
Published as open standards only on the first independent adopter (the Stage 4 gate on magician.network/roadmap), so the registry marks trust/1 and introduction/1 open:false with a publishGate and a machine learns they are held, not offered. A schema for each exists beside the vectors, derived from the parser and pinned to it, but is neither linked from the registry nor served.
Edge cases — each one paid for once
A declined introduction is indistinguishable from a path that never existed
toRequesterView collapses the two, and a test compares the objects for deep equality; a visible refusal would leak exactly the relationship data the refusal was protecting. Only the owner of an edge consents to crossing it — nothing constructs a consented hop directly, and there is no partial yes.
Nothing a person does alone moves their standing
Authority comes from what others assert; the measured figure counts only sealed non-none outcomes. There is a test whose whole job is proving unilateral activity moves nothing — a reward input would have to break it, which is the tripwire that keeps token mechanics out of trust scoring.