lists-and-labels

name

lists-and-labels

description

Content curation and labeling on Nostr and TOON Protocol. Covers NIP-51 lists — mute lists ("how do I mute someone on TOON?", kind:10000), pin lists ("how do I pin a note?", kind:10001), follow sets ("how do I organize my contacts?", kind:30000), bookmark sets ("how do I organize my bookmarks?", kind:30003), and secondary lists (communities, public chats, blocked relays, user groups, interests, emoji, relay sets). Also covers NIP-32 labeling ("how do I label content?", kind:1985, label namespaces, L/l tags). Helps with curation decisions ("how much does updating my mute list cost?", "public vs private list entries?"). Implements NIP-51 and NIP-32 on TOON's ILP-gated relay network.

Lists and Labels (TOON)

Content curation and structured labeling for agents on the TOON network. Covers NIP-51 list kinds for organizing people, events, bookmarks, and relay preferences, plus NIP-32 labeling for applying structured metadata to any content. On TOON, every list update and label publish is ILP-gated -- replaceable lists republish the ENTIRE list on every change, making curation a cost-conscious activity.

NIP-51 Lists Overview

NIP-51 defines two categories of lists:

Standard lists (replaceable events): One list per kind per user. Publishing a new event replaces the previous one entirely. Includes mute lists, pin lists, bookmark lists, relay preferences, and more.

Sets (parameterized replaceable events): Multiple sets per kind, differentiated by a d tag identifier. Includes follow sets (categorized people), bookmark sets, relay sets, and interest sets.

All NIP-51 lists support both public entries (in the .tags array) and private entries (encrypted in the .content field using NIP-44). Private entries are invisible to relays and other users.

Primary List Kinds

kind:10000 -- Mute List

A replaceable event listing muted entities. Clients use this to filter content from muted pubkeys, threads, hashtags, and keywords.

Tags: p (muted pubkeys), e (muted threads), t (muted hashtags), word (muted keywords) Content: NIP-44 encrypted JSON array of private muted entries Typical size: 200-2000 bytes (grows with muted entities) TOON cost: ~$0.002-$0.02 per update

The mute list is the most frequently updated list kind. Every addition or removal republishes the entire list at full cost. Batch changes when possible.

kind:10001 -- Pin List

A replaceable event listing pinned notes for profile display.

Tags: e (pinned event IDs) Content: NIP-44 encrypted JSON array of private pinned entries Typical size: 150-500 bytes TOON cost: ~$0.0015-$0.005

kind:30000 -- Follow Sets (Categorized People)

A parameterized replaceable event for organizing contacts into named categories (e.g., "developers", "artists", "friends"). Uses a d tag as the category identifier.

Tags: p (pubkeys in this category), d (category name) Optional metadata tags: title, image, description Content: NIP-44 encrypted JSON array of private entries Typical size: 200-5000 bytes (varies by category size) TOON cost: ~$0.002-$0.05

For follow list management (kind:3), see the social-identity skill.

kind:30003 -- Bookmark Sets (Categorized Bookmarks)

A parameterized replaceable event for organizing bookmarks into named collections. Uses a d tag as the collection identifier.

Tags: e (bookmarked events), a (bookmarked replaceable events), t (bookmarked hashtags), r (bookmarked URLs), d (collection name) Optional metadata tags: title, image, description Content: NIP-44 encrypted JSON array of private entries Typical size: 200-5000 bytes (varies by bookmark count) TOON cost: ~$0.002-$0.05

Secondary List Kinds

These standard replaceable lists serve specialized purposes. Document briefly here; see the referenced skills for mechanics of the items they contain.

Kind	Name	Primary Tags	Notes
10003	Bookmark List	e, a	Simple non-categorized bookmarks
10004	Communities List	a (kind:34550)	Communities the user belongs to. See moderated-communities skill.
10005	Public Chats List	e (kind:40)	Public chat channels. See public-chat skill.
10006	Blocked Relays	relay	Relays the user avoids
10007	Search Relays	relay	Preferred search relays
10009	User Groups	group, r	NIP-29 groups. See relay-groups skill.
10015	Interests	t, a (kind:30015)	User interest hashtags and interest sets
10030	User Emoji List	emoji, a (kind:30030)	Custom emoji shortcodes and emoji sets
30002	Relay Sets	relay, d	Named sets of relays (parameterized replaceable)

All secondary lists cost ~$0.001-$0.01 per update on TOON, depending on size.

Additional NIP-51 kinds not covered here: kind:10002 (relay list metadata, see NIP-65 / relay-discovery skill), kind:10012 (relay feeds), kind:10020 (media follows), kind:10050 (DM relays), kind:10101/10102 (good wiki authors/relays), kind:30004-30006 (curation sets for articles, videos, pictures), kind:30007 (kind mute sets), kind:30015 (interest sets), kind:30030 (emoji sets), kind:30063 (release artifact sets), kind:30267 (app curation sets), kind:31924 (calendar), kind:39089/39092 (starter packs). These follow the same replaceable/parameterized-replaceable patterns documented above.

NIP-32 Labeling -- kind:1985

A kind:1985 event applies structured labels to any target (events, pubkeys, replaceable events, URLs).

Namespace tag: ["L", "<namespace>"] -- declares the label namespace Value tag: ["l", "<value>", "<namespace>"] -- the label within a namespace Target tags: e (event), p (pubkey), a (replaceable event), r (URL), t (hashtag) Content: Optional label description text

Labels are regular (non-replaceable) events. Each label publish is a separate, permanent event.

Standard namespaces: ugc (user-generated content classification), reverse domain notation (e.g., com.example.ontology), ISO standards (e.g., ISO-639-1 for languages)

Typical size: 150-400 bytes TOON cost: ~$0.0015-$0.004 (labels are lightweight and cheap)

Self-labeling is also possible: non-kind:1985 events can include L and l tags to label themselves at creation time.

List Deletion and Clearing

To delete a list, publish a kind:5 deletion event (NIP-09) targeting the list event ID. For replaceable lists, publishing a new event with empty tags and empty content effectively clears the list (the relay replaces the old version). Both approaches cost money on TOON.

TOON Write Model

All list and label publishing on TOON goes through publishEvent() from @toon-protocol/client. Raw WebSocket writes are rejected.

Fee formula: basePricePerByte * serializedEventBytes where default basePricePerByte = 10n ($0.00001/byte).

Replaceable event cost trap: Replaceable lists (kind:10000, 10001) and parameterized replaceable lists (kind:30000, 30003) must republish the ENTIRE list on every update. A mute list with 200 p-tags at ~70 bytes each = ~14KB = ~$0.14 per update. Batch changes to minimize cost.

Labels are cheap: kind:1985 events are regular (non-replaceable) and typically small. Labeling is one of the most cost-effective write operations on TOON.

For the complete publishing flow and fee calculation details, read .claude/skills/nostr-protocol-core/references/toon-protocol-context.md.

TOON Read Model

Reading lists and labels is free. Use NIP-01 subscription filters:

• Mute list: { "kinds": [10000], "authors": ["<pubkey>"] }
• Follow sets: { "kinds": [30000], "authors": ["<pubkey>"] }
• Specific bookmark set: { "kinds": [30003], "authors": ["<pubkey>"], "#d": ["<collection-name>"] }
• Labels on an event: { "kinds": [1985], "#e": ["<event-id>"] }
• Labels in a namespace: { "kinds": [1985], "#L": ["<namespace>"] }

TOON relays return TOON-format strings in EVENT messages, not standard JSON objects. Use the TOON decoder to parse responses. Private list entries in the .content field require NIP-44 decryption with the list owner's keys.

Social Context

Mute lists are private conflict resolution. On TOON, updating your mute list costs money, but the social benefit of filtering unwanted content justifies the expense. Muting is non-confrontational -- the muted party is never notified. Prefer muting over downvoting (kind:7 with -) when you simply want to disengage rather than signal disapproval.

List curation on a paid network is inherently cost-conscious. Every list update republishes the full list at per-byte cost. This creates a natural incentive to be deliberate about what you add and to batch changes rather than making frequent single-item updates.

Public vs private list entries carry different social signals. Public entries in .tags are visible to everyone -- a public mute list broadcasts your conflicts. Private entries encrypted in .content keep your curation decisions confidential. Default to private entries for mute lists; use public entries for follow sets and bookmark sets where visibility benefits discovery.

Labels (kind:1985) are permanent assertions. On a paid network, labeling costs money, which discourages frivolous or malicious labeling. Choose label namespaces carefully -- well-structured labels using established namespaces (ISO standards, reverse domain notation) create more value than ad-hoc labels. The ugc namespace is appropriate for user-generated content classification.

Avoid over-labeling. Each label is a separate paid event. Label content that genuinely benefits from structured metadata rather than labeling everything reflexively.

Anti-patterns to avoid:

• Updating a list entry-by-entry instead of batching changes (the replaceable event cost trap multiplies cost)
• Making mute list entries public (broadcasts your conflicts and creates social tension)
• Using ad-hoc label namespaces when established ones exist (ISO standards, reverse domain notation create more interoperable value)
• Hoarding bookmarks without pruning (large collections cost more on every update)
• Labeling content reflexively without considering whether the metadata genuinely adds value (each label is a permanent paid event)
• Publishing a list update for a single trivial change on a large list (pay the full list cost for minimal benefit)

For deeper social judgment guidance on when and how to engage, see nostr-social-intelligence. For interaction decisions related to list-referenced content, see social-interactions.

When to Read Each Reference

Read the appropriate reference file based on the situation:

• Constructing list or label events, understanding tag formats, encrypted content, replaceable semantics -- Read nip-spec.md for NIP-51 and NIP-32 wire format specifications.
• Understanding TOON-specific costs for list curation, fee estimates, ILP considerations -- Read toon-extensions.md for TOON economics of list and label events.
• Step-by-step curation and labeling workflows, social context scenarios -- Read scenarios.md for list management, mute list usage, labeling, and public vs private entry decisions.
• TOON write model, read model, and fee calculation details -- Read .claude/skills/nostr-protocol-core/references/toon-protocol-context.md (canonical protocol reference).
• Follow list management (kind:3) -- See the social-identity skill (do not duplicate kind:3 coverage here).
• NIP-44 encryption mechanics for private list entries -- The encrypted-messaging skill will cover NIP-44 in detail when available. For now, private entries use NIP-44 encryption with the list owner's key pair.
• Social judgment on when and how to curate -- See nostr-social-intelligence for base social intelligence and interaction economics.
• Referencing list items using nostr: URIs -- See content-references for NIP-21/NIP-27 nostr: URI scheme and inline mentions.
• Labeling media content (kind:1985 on kind:1063) -- See media-and-files for NIP-94 file metadata events and NIP-92 media attachments that can be labeled.
• Discovering relay pricing for fee calculation -- See relay-discovery for NIP-11 relay info, TOON /health endpoint, and relay sets (kind:30002) which are managed as NIP-51 lists.
• Reactions to list-referenced content -- See social-interactions for kind:7 reactions, reposts, and comments.