You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Current anti-spam measures (IP rate limiting, basic PoW) are insufficient against coordinated Sybil attacks. Operators need an intelligent way to filter out spammers based on decentralized cryptographic trust networks.
Proposed Solution
Build an active spam filter driven by NIP-85 assertions from configured Trust Providers.
Scope
Assertion Ingestion: Will build a Trusted Assertion Service to ingest and cache kind: 30382 (User Assertions) from configured Trust Provider pubkeys.
Trust Score Caching: Will maintain a high-performance Redis/PostgreSQL cache mapping pubkey -> trust_score.
Spam Rejection Middleware: Will inject a validation step into the core event publishing pipeline to instantly reject events from negatively scored pubkeys (blocked: pubkey untrusted).
Not in the rewrite's scope, and the issue says why by itself: this proposes a new engine, not parity. A Trusted Assertion Service ingesting kind: 30382, a pubkey -> trust_score cache, and a rejection step injected into the publish path are three new components the TypeScript capture has no counterpart for — and NIP-85 is not among the 25 NIPs the relay advertises at the capture point, so it is neither parity work nor a Phase 0 audit gap.
Left without a milestone: the rewrite set is the work #860 owns, and a "Post-cutover" milestone would be a phase with no phase. Recorded in the "Deferred past cutover" section of the rewrite's estimates.md so the decision is not lost.
Worth carrying to whoever picks it up after cutover: #866 already ports the Web-of-Trust graph service from the capture (Redis-backed adjacency + the publish path), so this would be a consumer of that graph rather than a second one.
Activity
chappie-daemon commentedon Oct 24, 2026
Decision (2026-10-24): post-cutover
Not in the rewrite's scope, and the issue says why by itself: this proposes a new engine, not parity. A Trusted Assertion Service ingesting
kind: 30382, apubkey -> trust_scorecache, and a rejection step injected into the publish path are three new components the TypeScript capture has no counterpart for — and NIP-85 is not among the 25 NIPs the relay advertises at the capture point, so it is neither parity work nor a Phase 0 audit gap.Left without a milestone: the rewrite set is the work #860 owns, and a "Post-cutover" milestone would be a phase with no phase. Recorded in the "Deferred past cutover" section of the rewrite's
estimates.mdso the decision is not lost.Worth carrying to whoever picks it up after cutover: #866 already ports the Web-of-Trust graph service from the capture (Redis-backed adjacency + the publish path), so this would be a consumer of that graph rather than a second one.