# Trust Signals: Advisories, Attestations & Lineage

Beyond the automatic trust grade, ChainRender has a layer of **social trust signals** — ways for the community to
attest to, advise on, and trace the lineage of components. These are informational and ownerless; they inform
users without gating anything.

## Attestations (vouching)
An ownerless **attester registry** lets attesters vouch for components (genesis attesters + multi-hop vouches). An
attested dependency can lift a component's grade from unproven (C) to attested (B) — a verifiable statement that
someone staked their reputation on it. Trust is delegated, not centralized: you choose which attesters you trust.

## Advisories (alerting, not vouching)
An **advisory registry** is the inverse: a way to flag a *concern* about a component (e.g. "this resolver's source
changed"). Advisories *alert*, they don't *vouch* — and they never silently block a component. They surface
information so a user can decide, preserving the ownerless, un-censorable property.

## App DNA (forking & lineage)
Every composition can be **forked**, and a fork records its **parent on-chain**. This creates a verifiable lineage
— you can trace any app back through its ancestry. It makes ChainRender apps genuinely composable building blocks:
fork a known-good app, modify it, and its provenance is permanent and public.

## Annotations (component metadata)

The **`ComponentAnnotationRegistry`** is an ownerless, append-only key-value metadata layer for registered
components — a way to attach labels, descriptions, tags, or references to a component hash. Like everything in the
trust layer, an annotation is a **claim by its annotator**, keyed to a verified hash — a hint, not gospel. Multiple
annotators can annotate the same component; consumers weigh them by who they trust. It's how components carry
discoverable metadata (a name, a category, links) without any central catalog owning that data.

## Why these are separate from the grade

The grade measures *structural* trust (is it on-chain? are deps immutable?). These signals add *social* trust
(has anyone vouched? flagged? where did it come from?). Keeping them separate keeps the grade objective and
computable, while letting human judgment layer on top — transparently, ownerlessly, and without any party able to
censor or force a verdict.

## Related
- **[The Trust Grade](./trust-grade.md)** — the automatic, structural grade.
- **[Contracts → Registries](../contracts/registries.md)** — the on-chain registries.