@indexnetwork/protocol 66.0.0-rc.635.1 → 66.0.0-rc.638.1

package/README.md

@@ -12,7 +12,7 @@ Normative terms such as **MUST**, **MUST NOT**, **SHOULD**, and **MAY** are used
 
 Index Network is a private, intent-driven discovery protocol for agent-mediated opportunity discovery.
 
-Participants express **signals**: structured statements of what they seek, offer, are open to, or can credibly support. Agents interpret those signals against participant **context**, constrain discovery to bounded **communities**, evaluate candidate overlaps, negotiate fit when appropriate, and surface **opportunities** only through consent-gated state transitions.
+Participants express **intents**: structured statements of what they seek, offer, are open to, or can credibly support. Agents interpret those intents against participant **context**, constrain discovery to bounded **communities**, evaluate candidate overlaps, negotiate fit when appropriate, and surface **opportunities** only through consent-gated state transitions.
 
 The protocol is designed for high-signal human and agent coordination. It is not a public people database, keyword search engine, advertising channel, or automated introduction machine. Its purpose is to discover meaningful overlap while preserving context, scope, and human approval.
 
@@ -22,10 +22,10 @@ The protocol is designed for high-signal human and agent coordination. It is not
 flowchart LR
     Participant[Participant] --> Agent[Authorized agent]
     Participant --> Context[Context]
-    Participant --> Signal[Signal]
+    Participant --> Intent[Intent]
 
     Context --> Scope[Effective scope]
-    Signal --> Scope
+    Intent --> Scope
     Community[Community membership and norms] --> Scope
     AgentScope[Agent permissions] --> Scope
 
@@ -37,7 +37,7 @@ flowchart LR
     Negotiation -->|fit established| Opportunity
     Negotiation -->|human judgment required| Questions[Structured questions]
     Questions --> Participant
-    Participant --> Signal
+    Participant --> Intent
 
     Opportunity --> Consent{Participant consent?}
     Consent -->|yes| Connection[Connection]
@@ -48,8 +48,8 @@ The graph is intentionally consent-centered: agents can construct context, disco
 
 ## Design goals
 
-1. **Intent as the primary primitive** — discovery begins from a participant's current signal, not from static identity alone.
-2. **Bounded visibility** — communities define the scope in which signals and context may be evaluated.
+1. **Intent as the primary primitive** — discovery begins from a participant's current intent, not from static identity alone.
+2. **Bounded visibility** — communities define the scope in which intents and context may be evaluated.
 3. **Semantic discovery** — matching is based on role fit, constraints, complementarity, and contextual relevance rather than exact keyword overlap.
 4. **Explainable surfacing** — every surfaced opportunity SHOULD include a legible reason: why these participants, why now, and what the next action might be.
 5. **Consent at relationship boundaries** — agents MAY discover and negotiate, but MUST NOT create or accept a relationship without explicit participant approval.
@@ -72,7 +72,7 @@ The protocol does not attempt to be:
 |---|---|
 | **Participant** | A human principal represented in the network. A participant may act directly or through one or more agents. |
 | **Agent** | A software actor authorized to act for a participant or community within a declared scope. |
-| **Signal** | A participant's actionable expression of intent: what they seek, offer, need, are building, are exploring, or can support. |
+| **Intent** | A participant's actionable expression of intent: what they seek, offer, need, are building, are exploring, or can support. |
 | **Context** | A synthesized representation of a participant's identity, history, constraints, and community-specific relevance. |
 | **Community** | A bounded discovery scope with membership, purpose, norms, and relevance criteria. |
 | **Membership** | The relationship between a participant and a community. Agents receive community authority through separate scoped permissions. |
@@ -81,22 +81,22 @@ The protocol does not attempt to be:
 | **Negotiation** | A bounded agent-to-agent exchange used to test fit, constraints, timing, or consent before surfacing or advancing an opportunity. |
 | **Connection** | A participant-approved communication channel or introduction resulting from an accepted opportunity. |
 
-Some implementation APIs may expose historical names such as `intent`, `index`, `network`, `latent`, `pending`, `accepted`, `rejected`, `negotiating`, or `stalled`. Public-facing agents SHOULD translate stable participant states into protocol terms such as **signal**, **community**, **draft**, **sent**, **connected**, and **declined**, while treating negotiating or stalled states as internal process state unless the participant needs to act on them.
+Some implementation APIs may expose historical names such as `intent`, `index`, `network`, `latent`, `pending`, `accepted`, `rejected`, `negotiating`, or `stalled`. Public-facing agents SHOULD translate stable participant states into protocol terms such as **intent**, **community**, **draft**, **sent**, **connected**, and **declined**, while treating negotiating or stalled states as internal process state unless the participant needs to act on them.
 
 ## Object relationship model
 
 ```mermaid
 erDiagram
     PARTICIPANT ||--o{ AGENT : authorizes
-    PARTICIPANT ||--o{ SIGNAL : expresses
+    PARTICIPANT ||--o{ INTENT : expresses
     PARTICIPANT ||--o{ CONTEXT : represented_by
     PARTICIPANT ||--o{ MEMBERSHIP : holds
     AGENT ||--o{ AGENT_PERMISSION : receives
     COMMUNITY ||--o{ MEMBERSHIP : contains
     COMMUNITY ||--o{ AGENT_PERMISSION : scopes
-    COMMUNITY ||--o{ SIGNAL : scopes
+    COMMUNITY ||--o{ INTENT : scopes
     COMMUNITY ||--o{ CONTEXT : lenses
-    SIGNAL ||--o{ CANDIDATE : generates
+    INTENT ||--o{ CANDIDATE : generates
     CONTEXT ||--o{ CANDIDATE : generates
     CANDIDATE ||--o| OPPORTUNITY : promotes_to
     OPPORTUNITY ||--o{ NEGOTIATION : may_require
@@ -114,7 +114,7 @@ erDiagram
       string purpose
       string membership_policy
     }
-    SIGNAL {
+    INTENT {
       string constraints
       string lifecycle_state
     }
@@ -138,7 +138,7 @@ The protocol distinguishes human principals from software actors.
 
 ### Scope
 
-All discovery is scope-bound. A protocol operation MUST resolve an effective scope before cross-participant discovery, candidate generation, or opportunity evaluation. Implementations MAY read the requesting participant's own signals or context while preparing that operation.
+All discovery is scope-bound. A protocol operation MUST resolve an effective scope before cross-participant discovery, candidate generation, or opportunity evaluation. Implementations MAY read the requesting participant's own intents or context while preparing that operation.
 
 A scope may include:
 
@@ -181,27 +181,27 @@ Internal records MAY contain IDs, embeddings, scores, and graph state. Participa
 
 ## Core objects
 
-### Signal
+### Intent
 
-A signal is an actionable statement of direction. It may represent a need, offer, collaboration interest, hiring intent, funding goal, research direction, introduction request, or other future-oriented constraint.
+An intent is an actionable statement of direction. It may represent a need, offer, collaboration interest, hiring intent, funding goal, research direction, introduction request, or other future-oriented constraint.
 
-A signal SHOULD contain enough specificity to support discovery. Underspecified signals SHOULD enter clarification before they are persisted or used for broad discovery.
+An intent SHOULD contain enough specificity to support discovery. Underspecified intents SHOULD enter clarification before they are persisted or used for broad discovery.
 
-Signals have the following conceptual lifecycle:
+Intents have the following conceptual lifecycle:
 
 | State | Meaning |
 |---|---|
 | **Proposed** | A participant or agent supplied raw intent-like input. |
 | **Clarifying** | The protocol requires additional constraints before discovery. |
-| **Active** | The signal is valid, scoped, and eligible for discovery. |
+| **Active** | The intent is valid, scoped, and eligible for discovery. |
 | **Updated** | The participant refined or replaced constraints. |
-| **Archived / expired** | The signal should no longer produce new opportunities. |
+| **Archived / expired** | The intent should no longer produce new opportunities. |
 
-A signal MUST NOT be treated as active if it is outside the participant's authority, obviously insincere, unsafe to act on, or too vague to evaluate.
+An intent MUST NOT be treated as active if it is outside the participant's authority, obviously insincere, unsafe to act on, or too vague to evaluate.
 
 ### Context
 
-Context describes a participant: background, role, current work, capability, location, affiliation, constraints, and declared preferences. It grounds discovery by determining whether a signal is plausible and which communities or counterparts are relevant.
+Context describes a participant: background, role, current work, capability, location, affiliation, constraints, and declared preferences. It grounds discovery by determining whether an intent is plausible and which communities or counterparts are relevant.
 
 Context MAY be global to a participant or specific to a community. Community-specific context SHOULD emphasize facts relevant to that community's purpose and suppress irrelevant detail.
 
@@ -209,7 +209,7 @@ Context updates SHOULD cause downstream discovery representations to refresh. St
 
 ### Community
 
-A community is a bounded discovery environment. It defines who can participate, what kinds of signals are relevant, and which discovery operations are legitimate.
+A community is a bounded discovery environment. It defines who can participate, what kinds of intents are relevant, and which discovery operations are legitimate.
 
 A community SHOULD have:
 
@@ -228,7 +228,7 @@ An opportunity is an evaluated overlap between participants. It is not merely a
 An opportunity SHOULD include:
 
 - participating roles,
-- the relevant signal or context on each side,
+- the relevant intent or context on each side,
 - a concise explanation of fit,
 - a recommended next action,
 - lifecycle state,
@@ -293,8 +293,8 @@ sequenceDiagram
     participant C as Community scope
     participant N as Counterparty agent
 
-    P->>A: Provide context and signal
-    A->>PR: Admit signal and construct context
+    P->>A: Provide context and intent
+    A->>PR: Admit intent and construct context
     PR->>C: Resolve effective scope
     C-->>PR: Permitted communities
     PR->>PR: Generate candidates inside scope
@@ -307,7 +307,7 @@ sequenceDiagram
         PR-->>A: Draft opportunity or structured question
     end
     A->>P: Explain strongest reason and ask for approval
-    P-->>A: Approve, decline, or refine signal
+    P-->>A: Approve, decline, or refine intent
     A->>PR: Advance lifecycle only if approved
 ```
 
@@ -315,17 +315,17 @@ sequenceDiagram
 
 The protocol collects participant-provided or participant-authorized material and turns it into context. Context construction MUST preserve provenance and SHOULD prefer participant-approved information over inferred information.
 
-### 2. Signal admission
+### 2. Intent admission
 
-The protocol evaluates a proposed signal for specificity, sincerity, authority, and safety. If the signal is too broad, ambiguous, or missing critical constraints, the agent SHOULD ask focused clarification questions before running discovery.
+The protocol evaluates a proposed intent for specificity, sincerity, authority, and safety. If the intent is too broad, ambiguous, or missing critical constraints, the agent SHOULD ask focused clarification questions before running discovery.
 
 ### 3. Scope resolution
 
-The protocol determines the effective communities in which the signal can operate. Scope resolution MUST intersect participant membership, agent permissions, and request-time constraints. If the intersection is empty, discovery MUST NOT proceed.
+The protocol determines the effective communities in which the intent can operate. Scope resolution MUST intersect participant membership, agent permissions, and request-time constraints. If the intersection is empty, discovery MUST NOT proceed.
 
 ### 4. Candidate generation
 
-The protocol generates candidates by comparing signals and context inside the effective scope. Candidate generation MAY use multiple strategies, including signal-to-signal, context-to-signal, semantic retrieval, and directed target construction.
+The protocol generates candidates by comparing intents and context inside the effective scope. Candidate generation MAY use multiple strategies, including intent-to-intent, context-to-intent, semantic retrieval, and directed target construction.
 
 Candidate generation is not surfacing. Candidate data MUST remain internal until evaluation and visibility checks pass.
 
@@ -399,7 +399,7 @@ flowchart LR
     AgentGate --> ScopedDeps[Scoped protocol dependencies]
     ScopedDeps --> Runtime
 
-    Runtime --> Graphs[Discovery, context, signal graphs]
+    Runtime --> Graphs[Discovery, context, intent graphs]
     Graphs --> Results[Bounded results]
     Results --> Runtime
     Runtime --> ParticipantOutput[Participant-facing output rules]

package/STABILITY.md

@@ -46,7 +46,7 @@ Covered by SemVer below. Breaking changes require a **major** bump.
 | **Public API** | Model config helpers, request scope helpers (`deriveAllowedNetworkIds`, `deriveDiscoveryNetworkIds`), `requestContext`. |
 | **Interfaces** | Every port you implement to inject infrastructure (databases, embedder, cache, scraper, integration, …). |
 | **Shared schemas** | Zod schemas + inferred types that cross the boundary (underspecification, identity, network-assignment, chat-context, …). |
-| **Intents** | `Intents` — the whole signal capability as one class (lifecycle graph, verification, preparation) plus `IntentsDeps` and the `Prepare*` / `RecoveryField*` types. Replaced the six separate intent exports in 18.0.0. |
+| **Intents** | `Intents` — the whole intent capability as one class (lifecycle graph, verification, preparation) plus `IntentsDeps` and the `Prepare*` / `RecoveryField*` types. Replaced the six separate intent exports in 18.0.0. |
 | **Agents** | Structured LLM agents (`OpportunityEvaluator`, …). |
 
 `Negotiations`, `NegotiationDatabase`, the opening and turn decision functions,

package/dist/capabilities/intents.d.ts

@@ -6,7 +6,7 @@
  * implementation, grouped by what they do rather than by layer:
  *
  *   graph/               the lifecycle graph — prepare/create; infer/verify explicit updates
- *   intent.inferrer      an utterance into candidate signals
+ *   intent.inferrer      an utterance into candidate intents
  *   intent.verifier      felicity and entropy verdicts
  *   intent.preparer        a typed payload into an admitted draft or recovery form
  *
@@ -30,9 +30,9 @@ export type { PrepareAnswer, PrepareInput, PrepareResult, RecoveryField, Recover
  * {@link Intents.createGraph} is the one method that requires `database`.
  */
 export interface IntentsDeps {
-    /** Signal persistence. Required by {@link Intents.createGraph}. */
+    /** Intent persistence. Required by {@link Intents.createGraph}. */
     database?: IntentGraphDatabase;
-    /** Embedding generator used to vectorize executed signals. */
+    /** Embedding generator used to vectorize executed intents. */
     embedder?: EmbeddingGenerator;
     /** Host follow-up work started after a persist. */
     followUp?: IntentFollowUp;
@@ -620,7 +620,7 @@ export declare class Intents {
         flags: string[];
     }>;
     /**
-     * Prepare one signal draft for creation admission.
+     * Prepare one intent draft for creation admission.
      *
      * Fold recovery answers into the draft, then apply creation admission. A
      * ready result authorizes final review; needs_revision returns admission

package/dist/capabilities/intents.js

@@ -6,7 +6,7 @@
  * implementation, grouped by what they do rather than by layer:
  *
  *   graph/               the lifecycle graph — prepare/create; infer/verify explicit updates
- *   intent.inferrer      an utterance into candidate signals
+ *   intent.inferrer      an utterance into candidate intents
  *   intent.verifier      felicity and entropy verdicts
  *   intent.preparer        a typed payload into an admitted draft or recovery form
  *
@@ -56,7 +56,7 @@ export class Intents {
     }
     // ── Preparation ─────────────────────────────────────────────────────────────
     /**
-     * Prepare one signal draft for creation admission.
+     * Prepare one intent draft for creation admission.
      *
      * Fold recovery answers into the draft, then apply creation admission. A
      * ready result authorizes final review; needs_revision returns admission

package/dist/capabilities/networks.d.ts

@@ -9,7 +9,7 @@
  *   network.state        that graph's channel state
  *   membership.graph     the roster graph — add, list, remove members
  *   membership.state     that graph's channel state
- *   indexer.graph        signal↔community assignment, direct or model-evaluated
+ *   indexer.graph        intent↔community assignment, direct or model-evaluated
  *   indexer.state        that graph's channel state
  *   network.recommender  ranking public communities during onboarding
  *
@@ -885,7 +885,7 @@ export declare class Networks {
         };
     }, unknown, unknown, []>;
     /**
-     * Build the signal↔community assignment graph — link a signal to a community
+     * Build the intent↔community assignment graph — link an intent to a community
      * and unlink it.
      *
      * @throws If the instance was constructed without a `database`.

package/dist/capabilities/networks.js

@@ -9,7 +9,7 @@
  *   network.state        that graph's channel state
  *   membership.graph     the roster graph — add, list, remove members
  *   membership.state     that graph's channel state
- *   indexer.graph        signal↔community assignment, direct or model-evaluated
+ *   indexer.graph        intent↔community assignment, direct or model-evaluated
  *   indexer.state        that graph's channel state
  *   network.recommender  ranking public communities during onboarding
  *
@@ -50,9 +50,9 @@ export class Networks {
     createMembershipGraph() {
         return new NetworkMembershipGraphFactory(this.database("createMembershipGraph")).createGraph();
     }
-    // ── Signal assignment ───────────────────────────────────────────────────────
+    // ── Intent assignment ───────────────────────────────────────────────────────
     /**
-     * Build the signal↔community assignment graph — link a signal to a community
+     * Build the intent↔community assignment graph — link an intent to a community
      * and unlink it.
      *
      * @throws If the instance was constructed without a `database`.

package/dist/index.js

@@ -29,7 +29,7 @@ export { CANONICAL_GUIDANCE_SUMMARY, CANONICAL_GUIDANCE_TOPICS, CANONICAL_GUIDAN
 export { decideNegotiationOpening, decideNegotiationTurn, observeNegotiation, negotiationTurnSchema } from './protocol/negotiation.rules.js';
 // ─── Networks ─────────────────────────────────────────────────────────────────
 // The whole capability behind one class: the community lifecycle graph, the
-// membership graph, and signal assignment.
+// membership graph, and intent assignment.
 export { Networks } from "./capabilities/networks.js";
 export { NetworkPulseSummarizer } from "./internal/networks/network.pulse-summarizer.js";
 // ─── Intents ──────────────────────────────────────────────────────────────────

package/dist/internal/intents/graph/intent.graph.reconcile.d.ts

@@ -1,5 +1,5 @@
 /**
- * Intent graph, stages 2-3: verify inferred signals, then reconcile against what exists.
+ * Intent graph, stages 2-3: verify inferred intents, then reconcile against what exists.
  */
 import { VerifiedIntent, type IntentValidationFailure } from "./intent.graph.state.js";
 import type { DebugMetaAgent } from "../../../protocol/core.js";

package/dist/internal/intents/graph/intent.graph.reconcile.js

@@ -1,5 +1,5 @@
 /**
- * Intent graph, stages 2-3: verify inferred signals, then reconcile against what exists.
+ * Intent graph, stages 2-3: verify inferred intents, then reconcile against what exists.
  */
 import { admissionFailure } from "../intent.admission.js";
 import { timed } from "../../shared/observability/performance.js";

package/dist/internal/intents/intent.admission.js

@@ -2,7 +2,7 @@ const MAX_PERMISSIBLE_ENTROPY = 0.75;
 const MIN_CLEAR_INTENT_SCORE = 40;
 // Match the whole request so qualified job searches use the verifier's verdict.
 const GENERIC_JOB_PHRASE = /^(?:(?:i(?:['’]m|\s+am)?\s+)?(?:want|need|seek(?:ing)?|looking\s+for)\s+)?(?:a|any|some)\s+job[.!?]*$/i;
-const DEFAULT_SPECIFICITY_WARNING = "This signal is broad and may produce many weak matches. Add a more concrete role, outcome, location, timeframe, domain, or specific need to get better recommendations.";
+const DEFAULT_SPECIFICITY_WARNING = "This intent is broad and may produce many weak matches. Add a more concrete role, outcome, location, timeframe, domain, or specific need to get better recommendations.";
 /** The admission policy shared by preparation and explicit updates. */
 export function admissionFailure(description, verdict, isExplicitUpdate = false) {
     const details = { classification: verdict.classification, referentialBreadth: verdict.referential_breadth };

package/dist/internal/intents/intent.preparer.js

@@ -53,7 +53,7 @@ export class IntentPreparer {
         if (answers.length > 0) {
             const model = createStructuredModel("intentClarifier", payloadSchema, { name: "intent_draft" });
             const result = await invokeWithAbortSignal(model, [
-                new SystemMessage("Fold every supplied answer into the signal in the person's own voice. Preserve all existing details unless an answer explicitly changes them. Every claim must trace to the draft or answers; never invent constraints or facts to make a signal pass validation. Return the complete draft, concise and concrete."),
+                new SystemMessage("Fold every supplied answer into the intent in the person's own voice. Preserve all existing details unless an answer explicitly changes them. Every claim must trace to the draft or answers; never invent constraints or facts to make an intent pass validation. Return the complete draft, concise and concrete."),
                 new HumanMessage(JSON.stringify({ payload, answers })),
             ]);
             payload = payloadSchema.parse(result).payload;
@@ -64,7 +64,7 @@ export class IntentPreparer {
             return { status: "ready", payload, metadata: semanticMetadata(verdict) };
         const model = createStructuredModel("intentClarifier", recoverySchema, { name: "intent_recovery" });
         const result = await invokeWithAbortSignal(model, [
-            new SystemMessage("Build one recovery form that helps the person concretize their signal. Return 2-6 fields. Each field needs a draft-specific label, a kind (single, multi, or text), and for single/multi 2-5 concrete options anchored to the draft. Text fields need a placeholder. Never re-ask what the draft or answers already settle. Options must be matchable and specific. Do not expose scores, classifications, JSON, or internal vocabulary."),
+            new SystemMessage("Build one recovery form that helps the person concretize their intent. Return 2-6 fields. Each field needs a draft-specific label, a kind (single, multi, or text), and for single/multi 2-5 concrete options anchored to the draft. Text fields need a placeholder. Never re-ask what the draft or answers already settle. Options must be matchable and specific. Do not expose scores, classifications, JSON, or internal vocabulary."),
             new HumanMessage(JSON.stringify({
                 payload,
                 answers,

package/dist/internal/networks/indexer.graph.d.ts

@@ -9,7 +9,7 @@ import { IntentNetworkGraphState, type AssignmentResult } from "./indexer.state.
  * - read: List intent–network links (by intentId or by networkId)
  * - delete: Unlink an intent from a network
  *
- * ## Signal assignment policy
+ * ## Intent assignment policy
  *
  * There is none to apply: a link exists because its owner asked for it. The
  * assign node writes the row at score 1 with `mode: manual_override`.

package/dist/internal/networks/indexer.graph.js

@@ -183,7 +183,7 @@ export async function readNode(state, deps) {
                         })),
                         count: intents.length,
                         mode: "intents_in_network",
-                        note: "Community titles and full signal details come from the community and signal reads.",
+                        note: "Community titles and full intent details come from the community and intent reads.",
                     },
                 };
             }
@@ -200,7 +200,7 @@ export async function readNode(state, deps) {
                     })),
                     count: intents.length,
                     mode: "intents_in_network",
-                    note: "Community titles and full signal details come from the community and signal reads.",
+                    note: "Community titles and full intent details come from the community and intent reads.",
                 },
             };
         }

package/dist/internal/networks/indexer.state.d.ts

@@ -13,7 +13,7 @@ export interface AssignmentResult {
  * Intent Index Graph State.
  * Handles CRUD for the intent_networks junction table (linking intents to networks).
  *
- * ## Signal assignment policy
+ * ## Intent assignment policy
  *
  * There is none: a link exists because its owner asked for it, so the assign
  * node writes the row at score 1 with mode `manual_override`. No LLM call.

package/dist/internal/networks/indexer.state.js

@@ -3,7 +3,7 @@ import { Annotation } from "@langchain/langgraph";
  * Intent Index Graph State.
  * Handles CRUD for the intent_networks junction table (linking intents to networks).
  *
- * ## Signal assignment policy
+ * ## Intent assignment policy
  *
  * There is none: a link exists because its owner asked for it, so the assign
  * node writes the row at score 1 with mode `manual_override`. No LLM call.

package/dist/internal/networks/network.pulse-summarizer.d.ts

@@ -12,9 +12,9 @@ export declare const NetworkPulseSummaryOutputSchema: z.ZodObject<{
 export type NetworkPulseSummaryOutput = z.infer<typeof NetworkPulseSummaryOutputSchema>;
 export interface NetworkPulseSummaryInput {
     networkTitle: string;
-    /** Recent open signals in the network, newest first, as members wrote them. */
-    signals: string[];
-    /** Recent negotiations: the two signals each pair of agents is weighing. */
+    /** Recent open intents in the network, newest first, as members wrote them. */
+    intents: string[];
+    /** Recent negotiations: the two intents each pair of agents is weighing. */
     negotiations?: Array<{
         a: string;
         b: string;
@@ -34,7 +34,7 @@ export declare class NetworkPulseSummarizer {
     private model;
     constructor();
     /**
-     * @param input - The network's title, its recent open signals and negotiating pairs.
+     * @param input - The network's title, its recent open intents and negotiating pairs.
      * @returns Both texts (each null when there is nothing suitable), or null when the call fails.
      */
     invoke(input: NetworkPulseSummaryInput): Promise<NetworkPulseSummary | null>;

package/dist/internal/networks/network.pulse-summarizer.js

@@ -30,13 +30,13 @@ You write the "what people are looking for" blurb on a network's public invite p
 Anyone holding the invite link reads it, including people who are not members.
 
 TASK:
-Read the members' open signals and write a short, clear overview, warm in tone, of what is
-going on inside, so a visitor gets a feel for the room. Group signals into themes, e.g.
+Read the members' open intents and write a short, clear overview, warm in tone, of what is
+going on inside, so a visitor gets a feel for the room. Group intents into themes, e.g.
 "A few people are looking for cofounders and early hires, others want to meet investors who
 get crypto and AI, and at least one person just wants coffee before the conference."
 
 PRIVACY (strict):
-- Describe themes in aggregate. Never quote a signal or reuse its distinctive wording.
+- Describe themes in aggregate. Never quote an intent or reuse its distinctive wording.
 - No names of people, companies, products, projects, schools, or events.
 - No places more specific than a broad region, no dates, numbers, amounts, links, or handles.
 - Use loose quantities ("a few", "some", "several", "others"). Never "one person" for
@@ -50,12 +50,12 @@ STYLE:
   touch of warmth is welcome in one phrase (the range of the room, an unexpected neighbour of
   asks), but no punchlines and nothing at anyone's expense.
 - No emoji, no exclamation marks, no em dashes, no headings or lists.
-- Do not mention signals, agents, Index, or the network's name. Start with the content.
+- Do not mention intents, agents, Index, or the network's name. Start with the content.
 
-If fewer than three signals are suitable after the rules above, return an empty summary.
+If fewer than three intents are suitable after the rules above, return an empty summary.
 
 NEGOTIATIONS:
-You may also get pairs of signals that two members' agents are negotiating over right now.
+You may also get pairs of intents that two members' agents are negotiating over right now.
 In "negotiations", say in one or two sentences, at most 45 words, what the agents are talking
 about at a very high level: the kind of deal, never its details ("a few are working out
 cofounder intros, others are trading notes on a speaking slot"). Same privacy rules, same
@@ -75,18 +75,18 @@ export class NetworkPulseSummarizer {
         });
     }
     /**
-     * @param input - The network's title, its recent open signals and negotiating pairs.
+     * @param input - The network's title, its recent open intents and negotiating pairs.
      * @returns Both texts (each null when there is nothing suitable), or null when the call fails.
      */
     async invoke(input) {
         const clean = (text) => text.replace(/\s+/g, " ").trim().slice(0, 400);
-        const signals = input.signals.map(clean).filter(Boolean);
+        const intents = input.intents.map(clean).filter(Boolean);
         const pairs = (input.negotiations ?? []).map((p) => ({ a: clean(p.a), b: clean(p.b) })).filter((p) => p.a && p.b);
-        if (signals.length < 3 && pairs.length < 2)
+        if (intents.length < 3 && pairs.length < 2)
             return { summary: null, negotiations: null };
         const prompt = [
             `## Network\n${input.networkTitle}`,
-            `## Open signals\n${signals.map((s) => `- ${s}`).join("\n") || "(none)"}`,
+            `## Open intents\n${intents.map((s) => `- ${s}`).join("\n") || "(none)"}`,
             `## Negotiating pairs\n${pairs.map((p) => `- ${p.a}  <->  ${p.b}`).join("\n") || "(none)"}`,
         ].join("\n\n");
         try {

package/dist/internal/opportunities/opportunity.actor.js

@@ -13,7 +13,7 @@ export function normalizeOpportunityActorIntent(value) {
         // persisted with `intent: "None"` in dev and every downstream reader took
         // it for a real binding — the counterparty seat was granted a
         // consultation (whose park writes `recipientIntentId` straight at the
-        // intents table) on a signal that does not exist.
+        // intents table) on an intent that does not exist.
         || normalized.toLowerCase() === 'none') {
         return undefined;
     }

package/dist/internal/opportunities/opportunity.cards.d.ts

@@ -60,9 +60,11 @@ export declare function presentOpportunityCard(deps: OpportunityCardsDeps, oppor
 }): Promise<OpportunityCard | null>;
 /**
  * List the viewer's opportunities in the given statuses as presented cards,
- * newest first, one card per counterpart. A not-a-fit (rejected) row never
- * takes that slot from another status for the same person. Cards whose
- * presentation fails are dropped.
+ * one card per counterpart. An accepted row keeps that person's slot over a
+ * newer pending or negotiating one, and a pending row this viewer already
+ * committed to keeps it over a newer unanswered one. Among the same rank,
+ * newest wins. A rejected or expired row never takes the slot from a live
+ * one. Cards whose presentation fails are dropped.
  *
  * @param deps - Database, cache, and presenter.
  * @param input - Viewer, statuses, optional network/intent scope, limit, cache bypass, skeleton mode.

package/dist/internal/opportunities/opportunity.cards.js

@@ -33,6 +33,20 @@ function counterpartUserIds(opportunity, viewerId) {
 function viewerHasCommitted(opportunity, viewerId) {
     return (opportunity.committedActorIds ?? []).includes(viewerId);
 }
+/**
+ * Which row keeps a person's single radar slot.
+ * Accepted first, then a pending row this viewer already committed to,
+ * then other live rows. Rejected and expired never take the slot from those.
+ */
+function counterpartSlotRank(opportunity, viewerId) {
+    if (opportunity.status === 'rejected' || opportunity.status === 'expired')
+        return 3;
+    if (opportunity.status === 'accepted')
+        return 0;
+    if (opportunity.status === 'pending' && viewerHasCommitted(opportunity, viewerId))
+        return 1;
+    return 2;
+}
 function stampViewerCommitted(card, viewerCommitted) {
     return { ...card, viewerCommitted };
 }
@@ -113,9 +127,11 @@ export async function presentOpportunityCard(deps, opportunity, viewerId, option
 }
 /**
  * List the viewer's opportunities in the given statuses as presented cards,
- * newest first, one card per counterpart. A not-a-fit (rejected) row never
- * takes that slot from another status for the same person. Cards whose
- * presentation fails are dropped.
+ * one card per counterpart. An accepted row keeps that person's slot over a
+ * newer pending or negotiating one, and a pending row this viewer already
+ * committed to keeps it over a newer unanswered one. Among the same rank,
+ * newest wins. A rejected or expired row never takes the slot from a live
+ * one. Cards whose presentation fails are dropped.
  *
  * @param deps - Database, cache, and presenter.
  * @param input - Viewer, statuses, optional network/intent scope, limit, cache bypass, skeleton mode.
@@ -133,9 +149,9 @@ export async function listOpportunityCards(deps, input) {
     const visible = raw
         .filter((opp) => requested.has(opp.status) && canUserSeeOpportunity(opp.actors, opp.status, viewerId))
         .sort((a, b) => {
-        const failed = Number(a.status === 'rejected') - Number(b.status === 'rejected');
-        if (failed !== 0)
-            return failed;
+        const rank = counterpartSlotRank(a, viewerId) - counterpartSlotRank(b, viewerId);
+        if (rank !== 0)
+            return rank;
         return b.createdAt.getTime() - a.createdAt.getTime();
     });
     const seen = new Set();

package/dist/internal/opportunities/opportunity.graph.modes.js

@@ -45,7 +45,7 @@ export async function readOpportunities(deps, request) {
                 return {
                     readResult: {
                         count: 0,
-                        message: 'You have no opportunities yet. Create or refine an approved signal; matching runs in the background. Review persisted results later from the opportunities read.',
+                        message: 'You have no opportunities yet. Create or refine an approved intent; matching runs in the background. Review persisted results later from the opportunities read.',
                         opportunities: [],
                     },
                 };

package/dist/internal/opportunities/opportunity.presentation.d.ts

@@ -101,7 +101,7 @@ export interface PresenterInput {
     matchReasoning: string;
     category: string;
     confidence: number;
-    signalsSummary: string;
+    intentsSummary: string;
     networkName: string;
     viewerRole: string;
     opportunityStatus?: string;
@@ -150,11 +150,11 @@ export declare class OpportunityPresenter {
     }): Promise<CardLLMResult[]>;
 }
 /**
- * Build the LLM-facing signal summary while excluding pool adjustments. Pool
+ * Build the LLM-facing intent summary while excluding pool adjustments. Pool
  * disposition is rendered deterministically by the card chip; asking the
  * presenter to interpret it could turn a demotion into a positive rationale.
  */
-export declare function summarizeSignalsForPresenter(signals: Opportunity['interpretation']['signals']): string;
+export declare function summarizeIntentsForPresenter(intents: Opportunity['interpretation']['intents']): string;
 /**
  * Gather all context needed for the presenter from the database.
  * Fetches viewer profile, viewer intents, other party profile(s), and network in parallel.