jev-recipes 0.7.0 → 0.8.0

package/README.md

@@ -11,7 +11,7 @@
 
 <!-- BEGIN GENERATED: summary -->
 
-243 focused recipes for JavaScript and TypeScript. Route messages, check evidence, and label model responses with a function call.
+246 focused recipes for JavaScript and TypeScript. Route messages, check evidence, and label model responses with a function call.
 
 <!-- END GENERATED: summary -->
 
@@ -19,6 +19,10 @@ Each recipe accepts your data, calls [Jev through TypeSafe's API](https://docs.t
 
 For example, give `route` a support message and descriptions of your teams. It returns a team such as `billing`, or a review outcome when the choice is uncertain.
 
+Evaluate your own cases with the [installed evaluator](https://github.com/agencyenterprise/jev-recipes/blob/main/docs/evaluation.md), retain model responses, and replay confidence policies offline. The [agent workflow](https://github.com/agencyenterprise/jev-recipes/blob/main/examples/agent-loop/README.md) and [customer queue](https://github.com/agencyenterprise/jev-recipes/blob/main/examples/customer-queue/README.md) show how decisions fit into application code. [Direct and Gateway integrations](https://github.com/agencyenterprise/jev-recipes/blob/main/docs/integrations.md) use the same recipe interface.
+
+Build the searchable static catalog with `npm run site:build`, then preview it with `npm run site:preview`. It includes fixture exploration, related-recipe comparisons, and saved evaluation evidence. Current measurements and missing or older evidence are labeled explicitly.
+
 ## Watch Jev play checkers against Jev using the checkers-move recipe
 
 [![Jev plays checkers against Jev using the checkers-move recipe](https://raw.githubusercontent.com/agencyenterprise/jev-recipes/main/examples/checkers/jev-vs-jev.gif)](https://www.youtube.com/shorts/Z282rGKysTg)

package/dist/catalog/generated/details/callback-responsibility.json

@@ -0,0 +1,45 @@
+{
+  "inputSchema": {
+    "$schema": "https://json-schema.org/draft/2020-12/schema",
+    "type": "object",
+    "properties": {
+      "conversation": { "type": "string" },
+      "roles": { "type": "string" },
+      "minConfidence": { "type": "number", "minimum": 0, "maximum": 1 }
+    },
+    "required": ["conversation", "roles"]
+  },
+  "resultSchema": {
+    "$schema": "https://json-schema.org/draft/2020-12/schema",
+    "type": "object",
+    "properties": {
+      "model": { "type": "string" },
+      "usage": {
+        "type": "object",
+        "properties": {
+          "input_tokens": { "type": "integer", "minimum": 0, "maximum": 9007199254740991 },
+          "output_tokens": { "type": "integer", "minimum": 0, "maximum": 9007199254740991 }
+        },
+        "required": ["input_tokens", "output_tokens"],
+        "additionalProperties": false
+      },
+      "status": { "type": "string", "enum": ["ready", "review"] },
+      "verdict": {
+        "type": "string",
+        "enum": ["business", "customer", "either", "none", "unclear"]
+      },
+      "confidence": { "type": "number", "minimum": 0, "maximum": 1 },
+      "probabilities": {
+        "type": "object",
+        "propertyNames": {
+          "type": "string",
+          "enum": ["business", "customer", "either", "none", "unclear"]
+        },
+        "additionalProperties": { "type": "number", "minimum": 0, "maximum": 1 },
+        "required": ["business", "customer", "either", "none", "unclear"]
+      }
+    },
+    "required": ["model", "usage", "status", "verdict", "confidence", "probabilities"],
+    "additionalProperties": false
+  }
+}

package/dist/catalog/generated/details/contact-opt-out.json

@@ -0,0 +1,45 @@
+{
+  "inputSchema": {
+    "$schema": "https://json-schema.org/draft/2020-12/schema",
+    "type": "object",
+    "properties": {
+      "message": { "type": "string" },
+      "context": { "type": "string" },
+      "minConfidence": { "type": "number", "minimum": 0, "maximum": 1 }
+    },
+    "required": ["message"]
+  },
+  "resultSchema": {
+    "$schema": "https://json-schema.org/draft/2020-12/schema",
+    "type": "object",
+    "properties": {
+      "model": { "type": "string" },
+      "usage": {
+        "type": "object",
+        "properties": {
+          "input_tokens": { "type": "integer", "minimum": 0, "maximum": 9007199254740991 },
+          "output_tokens": { "type": "integer", "minimum": 0, "maximum": 9007199254740991 }
+        },
+        "required": ["input_tokens", "output_tokens"],
+        "additionalProperties": false
+      },
+      "status": { "type": "string", "enum": ["ready", "review"] },
+      "verdict": {
+        "type": "string",
+        "enum": ["all_contact", "channel", "campaign", "none", "unclear"]
+      },
+      "confidence": { "type": "number", "minimum": 0, "maximum": 1 },
+      "probabilities": {
+        "type": "object",
+        "propertyNames": {
+          "type": "string",
+          "enum": ["all_contact", "channel", "campaign", "none", "unclear"]
+        },
+        "additionalProperties": { "type": "number", "minimum": 0, "maximum": 1 },
+        "required": ["all_contact", "channel", "campaign", "none", "unclear"]
+      }
+    },
+    "required": ["model", "usage", "status", "verdict", "confidence", "probabilities"],
+    "additionalProperties": false
+  }
+}

package/dist/catalog/generated/details/followup-timing.json

@@ -0,0 +1,45 @@
+{
+  "inputSchema": {
+    "$schema": "https://json-schema.org/draft/2020-12/schema",
+    "type": "object",
+    "properties": {
+      "message": { "type": "string" },
+      "context": { "type": "string" },
+      "minConfidence": { "type": "number", "minimum": 0, "maximum": 1 }
+    },
+    "required": ["message"]
+  },
+  "resultSchema": {
+    "$schema": "https://json-schema.org/draft/2020-12/schema",
+    "type": "object",
+    "properties": {
+      "model": { "type": "string" },
+      "usage": {
+        "type": "object",
+        "properties": {
+          "input_tokens": { "type": "integer", "minimum": 0, "maximum": 9007199254740991 },
+          "output_tokens": { "type": "integer", "minimum": 0, "maximum": 9007199254740991 }
+        },
+        "required": ["input_tokens", "output_tokens"],
+        "additionalProperties": false
+      },
+      "status": { "type": "string", "enum": ["ready", "review"] },
+      "verdict": {
+        "type": "string",
+        "enum": ["now", "later", "after_event", "not_requested", "declined", "unclear"]
+      },
+      "confidence": { "type": "number", "minimum": 0, "maximum": 1 },
+      "probabilities": {
+        "type": "object",
+        "propertyNames": {
+          "type": "string",
+          "enum": ["now", "later", "after_event", "not_requested", "declined", "unclear"]
+        },
+        "additionalProperties": { "type": "number", "minimum": 0, "maximum": 1 },
+        "required": ["now", "later", "after_event", "not_requested", "declined", "unclear"]
+      }
+    },
+    "required": ["model", "usage", "status", "verdict", "confidence", "probabilities"],
+    "additionalProperties": false
+  }
+}

package/dist/catalog/generated/details/game-action.json

@@ -3,101 +3,17 @@
     "$schema": "https://json-schema.org/draft/2020-12/schema",
     "type": "object",
     "properties": {
-      "gameState": {
-        "description": "The current game information available for this decision.",
-        "$ref": "#/$defs/__schema0"
-      },
-      "playerState": {
-        "description": "Information about the player making this decision.",
-        "$ref": "#/$defs/__schema1"
-      },
+      "gameState": { "description": "The current game information available for this decision." },
+      "playerState": { "description": "Information about the player making this decision." },
       "legalActions": {
         "type": "array",
-        "items": { "$ref": "#/$defs/__schema2" },
+        "items": {},
         "description": "The available actions the player can choose from."
       },
-      "rules": {
-        "description": "Game rules and explanations of game-specific information.",
-        "$ref": "#/$defs/__schema3"
-      },
-      "objective": {
-        "description": "What the player is trying to accomplish.",
-        "$ref": "#/$defs/__schema4"
-      }
+      "rules": { "description": "Game rules and explanations of game-specific information." },
+      "objective": { "description": "What the player is trying to accomplish." }
     },
-    "required": ["gameState", "playerState", "legalActions"],
-    "$defs": {
-      "__schema0": {
-        "anyOf": [
-          { "type": "string" },
-          { "type": "number" },
-          { "type": "boolean" },
-          { "type": "null" },
-          { "type": "array", "items": { "$ref": "#/$defs/__schema0" } },
-          {
-            "type": "object",
-            "propertyNames": { "type": "string" },
-            "additionalProperties": { "$ref": "#/$defs/__schema0" }
-          }
-        ]
-      },
-      "__schema1": {
-        "anyOf": [
-          { "type": "string" },
-          { "type": "number" },
-          { "type": "boolean" },
-          { "type": "null" },
-          { "type": "array", "items": { "$ref": "#/$defs/__schema1" } },
-          {
-            "type": "object",
-            "propertyNames": { "type": "string" },
-            "additionalProperties": { "$ref": "#/$defs/__schema1" }
-          }
-        ]
-      },
-      "__schema2": {
-        "anyOf": [
-          { "type": "string" },
-          { "type": "number" },
-          { "type": "boolean" },
-          { "type": "null" },
-          { "type": "array", "items": { "$ref": "#/$defs/__schema2" } },
-          {
-            "type": "object",
-            "propertyNames": { "type": "string" },
-            "additionalProperties": { "$ref": "#/$defs/__schema2" }
-          }
-        ]
-      },
-      "__schema3": {
-        "anyOf": [
-          { "type": "string" },
-          { "type": "number" },
-          { "type": "boolean" },
-          { "type": "null" },
-          { "type": "array", "items": { "$ref": "#/$defs/__schema3" } },
-          {
-            "type": "object",
-            "propertyNames": { "type": "string" },
-            "additionalProperties": { "$ref": "#/$defs/__schema3" }
-          }
-        ]
-      },
-      "__schema4": {
-        "anyOf": [
-          { "type": "string" },
-          { "type": "number" },
-          { "type": "boolean" },
-          { "type": "null" },
-          { "type": "array", "items": { "$ref": "#/$defs/__schema4" } },
-          {
-            "type": "object",
-            "propertyNames": { "type": "string" },
-            "additionalProperties": { "$ref": "#/$defs/__schema4" }
-          }
-        ]
-      }
-    }
+    "required": ["gameState", "playerState", "legalActions"]
   },
   "resultSchema": {
     "$schema": "https://json-schema.org/draft/2020-12/schema",
@@ -120,8 +36,7 @@
             "description": "The internal Choice label associated with the selected action."
           },
           "action": {
-            "description": "The original selected action, preserving any existing game ID.",
-            "$ref": "#/$defs/__schema0"
+            "description": "The original selected action, preserving any existing game ID."
           },
           "confidence": {
             "type": "number",
@@ -140,22 +55,6 @@
         "additionalProperties": false
       },
       { "type": "null" }
-    ],
-    "$defs": {
-      "__schema0": {
-        "anyOf": [
-          { "type": "string" },
-          { "type": "number" },
-          { "type": "boolean" },
-          { "type": "null" },
-          { "type": "array", "items": { "$ref": "#/$defs/__schema0" } },
-          {
-            "type": "object",
-            "propertyNames": { "type": "string" },
-            "additionalProperties": { "$ref": "#/$defs/__schema0" }
-          }
-        ]
-      }
-    }
+    ]
   }
 }

package/dist/catalog/generated/loaders.js

@@ -24,6 +24,7 @@ export const recipeLoaders = {
     'bug-report-completeness': () => import('../../recipes/bug-report-completeness/index.js').then((recipe) => (input, options) => recipe.bugReportCompleteness(recipe.bugReportCompletenessInputSchema.parse(input), options)),
     'buying-intent': () => import('../../recipes/buying-intent/index.js').then((recipe) => (input, options) => recipe.buyingIntent(recipe.buyingIntentInputSchema.parse(input), options)),
     'cache-match': () => import('../../recipes/cache-match/index.js').then((recipe) => (input, options) => recipe.cacheMatch(recipe.cacheMatchInputSchema.parse(input), options)),
+    'callback-responsibility': () => import('../../recipes/callback-responsibility/index.js').then((recipe) => (input, options) => recipe.callbackResponsibility(recipe.callbackResponsibilityInputSchema.parse(input), options)),
     'cancellation-check': () => import('../../recipes/cancellation-check/index.js').then((recipe) => (input, options) => recipe.cancellationCheck(recipe.cancellationCheckInputSchema.parse(input), options)),
     'care-urgency-wording': () => import('../../recipes/care-urgency-wording/index.js').then((recipe) => (input, options) => recipe.careUrgencyWording(recipe.careUrgencyWordingInputSchema.parse(input), options)),
     'category-fit': () => import('../../recipes/category-fit/index.js').then((recipe) => (input, options) => recipe.categoryFit(recipe.categoryFitInputSchema.parse(input), options)),
@@ -51,6 +52,7 @@ export const recipeLoaders = {
     'consent-request': () => import('../../recipes/consent-request/index.js').then((recipe) => (input, options) => recipe.consentRequest(recipe.consentRequestInputSchema.parse(input), options)),
     'consent-scope-fit': () => import('../../recipes/consent-scope-fit/index.js').then((recipe) => (input, options) => recipe.consentScopeFit(recipe.consentScopeFitInputSchema.parse(input), options)),
     'constraint-strength': () => import('../../recipes/constraint-strength/index.js').then((recipe) => (input, options) => recipe.constraintStrength(recipe.constraintStrengthInputSchema.parse(input), options)),
+    'contact-opt-out': () => import('../../recipes/contact-opt-out/index.js').then((recipe) => (input, options) => recipe.contactOptOut(recipe.contactOptOutInputSchema.parse(input), options)),
     'content-facets': () => import('../../recipes/content-facets/index.js').then((recipe) => (input, options) => recipe.contentFacets(recipe.contentFacetsInputSchema.parse(input), options)),
     'content-freshness-signal': () => import('../../recipes/content-freshness-signal/index.js').then((recipe) => (input, options) => recipe.contentFreshnessSignal(recipe.contentFreshnessSignalInputSchema.parse(input), options)),
     'context-prune': () => import('../../recipes/context-prune/index.js').then((recipe) => (input, options) => recipe.contextPrune(recipe.contextPruneInputSchema.parse(input), options)),
@@ -91,6 +93,7 @@ export const recipeLoaders = {
     'financial-advice-signal': () => import('../../recipes/financial-advice-signal/index.js').then((recipe) => (input, options) => recipe.financialAdviceSignal(recipe.financialAdviceSignalInputSchema.parse(input), options)),
     'finger-actions': () => import('../../recipes/finger-actions/index.js').then((recipe) => (input, options) => recipe.fingerActions(recipe.fingerActionsInputSchema.parse(input), options)),
     'followup-link': () => import('../../recipes/followup-link/index.js').then((recipe) => (input, options) => recipe.followupLink(recipe.followupLinkInputSchema.parse(input), options)),
+    'followup-timing': () => import('../../recipes/followup-timing/index.js').then((recipe) => (input, options) => recipe.followupTiming(recipe.followupTimingInputSchema.parse(input), options)),
     'forecast-confidence-wording': () => import('../../recipes/forecast-confidence-wording/index.js').then((recipe) => (input, options) => recipe.forecastConfidenceWording(recipe.forecastConfidenceWordingInputSchema.parse(input), options)),
     'format-fit': () => import('../../recipes/format-fit/index.js').then((recipe) => (input, options) => recipe.formatFit(recipe.formatFitInputSchema.parse(input), options)),
     'freshness-needed': () => import('../../recipes/freshness-needed/index.js').then((recipe) => (input, options) => recipe.freshnessNeeded(recipe.freshnessNeededInputSchema.parse(input), options)),

package/dist/catalog/generated/metadata.js

@@ -523,6 +523,30 @@ export const recipeMetadata = [
             },
         ],
     },
+    {
+        id: 'callback-responsibility',
+        title: 'Identify who should initiate a callback',
+        description: 'Which party is expected to initiate the next call?',
+        category: 'conversation',
+        tags: ['conversation', 'callback', 'ownership', 'handover'],
+        limitations: [
+            'Requires speaker roles. Does not infer identities from names or assume every agent speaks for the business.',
+            'Only identifies initiation of a future call, including a conditional call. Does not assign general task ownership or verify the call occurred.',
+            'An unaccepted suggestion, contradictory exchange, or unresolved pronoun requires review.',
+        ],
+        useWhen: 'You need to decide whether the business or customer is expected to initiate the next call.',
+        related: [
+            {
+                id: 'promise-check',
+                reason: 'Use promise-check to check whether a draft introduces an unsupported commitment.',
+            },
+            {
+                id: 'commitment-strength',
+                reason: 'Use commitment-strength to grade how firmly a statement commits its speaker.',
+            },
+            { id: 'followup-timing', reason: 'Use followup-timing for when another contact is wanted.' },
+        ],
+    },
     {
         id: 'cancellation-check',
         title: 'Check cancellation intent',
@@ -1161,6 +1185,33 @@ export const recipeMetadata = [
             },
         ],
     },
+    {
+        id: 'contact-opt-out',
+        title: 'Recognize a contact opt-out',
+        description: 'What scope of future contact does the sender ask to stop?',
+        category: 'conversation',
+        tags: ['conversation', 'contact', 'opt-out', 'preferences'],
+        limitations: [
+            'Reports the expressed scope category, not legal consent or permission to contact. A none result never grants consent.',
+            'Does not resolve channel or campaign identifiers or change subscription settings. Applications own those mappings and prior preferences.',
+            'Mixed channel and campaign restrictions, conflicting requests, and missing references require review.',
+        ],
+        useWhen: 'You need to distinguish stopping all contact from stopping a channel or campaign.',
+        related: [
+            {
+                id: 'cancellation-check',
+                reason: 'Use cancellation-check for cancelling a product or service, rather than future contact.',
+            },
+            {
+                id: 'buying-intent',
+                reason: 'Use buying-intent for purchase interest. Declining an offer alone does not establish an opt-out.',
+            },
+            {
+                id: 'followup-timing',
+                reason: 'Use followup-timing for a requested delay or condition for future contact.',
+            },
+        ],
+    },
     {
         id: 'content-facets',
         title: 'Label article facets',
@@ -2047,6 +2098,33 @@ export const recipeMetadata = [
             },
         ],
     },
+    {
+        id: 'followup-timing',
+        title: 'Recognize requested follow-up timing',
+        description: 'When does the sender want another contact, if any?',
+        category: 'conversation',
+        tags: ['conversation', 'followup', 'timing', 'scheduling'],
+        limitations: [
+            'Does not extract dates, time zones, or event identifiers and does not schedule contact.',
+            'A timing request does not override stored contact preferences or establish permission to contact.',
+            'Missing references, unresolved alternatives, and conflicting timing instructions require review.',
+        ],
+        useWhen: 'You need to distinguish immediate contact, a later time, an event condition, and no follow-up request.',
+        related: [
+            {
+                id: 'response-needed',
+                reason: 'Use response-needed to decide whether the current message needs a reply.',
+            },
+            {
+                id: 'contact-opt-out',
+                reason: 'Use contact-opt-out for the scope of a request to stop future contact.',
+            },
+            {
+                id: 'callback-responsibility',
+                reason: 'Use callback-responsibility to identify who should initiate a call.',
+            },
+        ],
+    },
     {
         id: 'forecast-confidence-wording',
         title: 'Grade the certainty of a financial projection',

package/dist/catalog/generated/names.d.ts

@@ -1 +1 @@
-export declare const recipeNames: readonly ["action-compare", "action-effects", "action-reversibility", "action-scope", "age-appropriateness", "alert-actionability", "answer-consistency", "answer-coverage", "answer-disclosures", "answer-grade", "answer-invalidation", "answer-relevance", "answerability", "appeal-grounds-kind", "appointment-request-kind", "argument-fit", "attempted-step", "attribution-match", "audience-fit", "breaking-change-signal", "budget-fit", "budget-narrative-fit", "bug-report-completeness", "buying-intent", "cache-match", "cancellation-check", "care-urgency-wording", "category-fit", "causal-attribution", "causal-language-strength", "certainty-match", "change-meaning", "change-risk", "change-window-fit", "checkers-move", "choose-action", "citation-match", "citation-needed", "claim-facets", "claim-stance", "clarify", "clause-conflict", "clause-kind", "clickbait-level", "commit-message-fit", "commitment-strength", "comparable-fit", "completion-gate", "confirmation-match", "consent-request", "consent-scope-fit", "constraint-strength", "content-facets", "content-freshness-signal", "context-prune", "context-role", "continuation-compare", "correction-target", "corrective-action-fit", "daw-request-kind", "deadline-risk", "defect-report-facets", "delegation-fit", "diff-hazards", "disclosure-facets", "dispute-kind", "document-role", "draft-compare", "dynamic-change", "eligibility-facets", "emotion-kind", "entity-match", "error-acknowledgment", "escalation-wording", "evaluation-mention", "evidence-conflict", "evidence-independence", "evidence-novelty", "evidence-strength", "exercise-select", "expense-category", "explanation-level", "extraction-fidelity", "fact-stability", "failure-kind", "fair-housing-wording", "feedback-actionability", "feedback-kind", "field-select", "financial-advice-signal", "finger-actions", "followup-link", "forecast-confidence-wording", "format-fit", "freshness-needed", "frustration-signal", "funder-fit", "game-action", "game-phase", "goal-drift", "grounding-level", "guest-request-facets", "handoff", "handoff-completeness", "headline-fit", "incident-match", "incident-severity-wording", "injection-signal", "instruction-clarity", "instruction-conflict", "instruction-fit", "instruction-priority", "instruction-readability", "instrument-issue-kind", "instrument-report-facets", "intake-question-fit", "intent-change", "invoice-facets", "issue-impact", "issue-recurrence", "itinerary-feasibility", "job-post-facets", "length-fit", "listener-request-kind", "listing-compare", "listing-fact-consistency", "log-line-kind", "loss-cause-kind", "lyric-mood-fit", "maintenance-hazard-wording", "medication-mention", "memory-relation", "memory-scope", "memory-subject", "memory-value", "message-facets", "methods-facets", "model-route", "modulation-moment", "mood-match", "mood-request", "motivation-source", "move-explanation-fit", "narrative-consistency", "next-chord", "next-duration", "next-note", "notice-facets", "novelty-claim-level", "objection-kind", "objective-fit", "offer-terms-facets", "outcome-framing", "passage-compare", "passage-difficulty", "passage-duplicate", "passage-mood", "passage-standalone", "performance-feedback-facets", "persuasion-technique", "phrase-complete", "pii-presence", "plan-completeness", "policy-compliance", "policy-severity", "politeness-level", "postmortem-facets", "preference-kind", "priority-compare", "privacy-notice-facets", "product-match", "progress-stall", "promise-check", "proposal-facets", "qualification-evidence", "query-equivalence", "query-specificity", "question-assumption", "question-leading", "question-relevance", "reconciliation-match", "reference-resolve", "repeated-attempt", "repetition-level", "reply-template-match", "report-facets", "requirement-testability", "rerank", "resolution-check", "response-needed", "response-refusal", "result-outcome", "result-plausibility", "result-usefulness", "retrieval-needed", "retry-worthwhile", "review-comment-kind", "review-facets", "review-recommendation-kind", "review-response-fit", "rollback-signal", "root-cause-depth", "route", "route-many", "rule-compliance", "runbook-fit", "safety-incident-kind", "satisfaction-signal", "search-intent-kind", "sentiment-shift", "settlement-offer-facets", "shipment-issue-kind", "slot-fit", "sound-match", "source-applicability", "spam-signal", "stage-evidence", "step-complete", "step-progress", "style-kind", "summary-coverage", "symptom-facets", "take-turn", "task-complexity", "task-dependency", "task-duplicate", "task-overlap", "tempo-change", "tension-level", "ticket-match", "tone-check", "tool-call-gate", "tool-compare", "tool-fit", "topic-shift", "trip-purpose-kind", "troubleshooting-fit", "turn-intent", "uncertainty-expression", "urgency-signal", "verify", "wake-gate", "workaround-fit"];
+export declare const recipeNames: readonly ["action-compare", "action-effects", "action-reversibility", "action-scope", "age-appropriateness", "alert-actionability", "answer-consistency", "answer-coverage", "answer-disclosures", "answer-grade", "answer-invalidation", "answer-relevance", "answerability", "appeal-grounds-kind", "appointment-request-kind", "argument-fit", "attempted-step", "attribution-match", "audience-fit", "breaking-change-signal", "budget-fit", "budget-narrative-fit", "bug-report-completeness", "buying-intent", "cache-match", "callback-responsibility", "cancellation-check", "care-urgency-wording", "category-fit", "causal-attribution", "causal-language-strength", "certainty-match", "change-meaning", "change-risk", "change-window-fit", "checkers-move", "choose-action", "citation-match", "citation-needed", "claim-facets", "claim-stance", "clarify", "clause-conflict", "clause-kind", "clickbait-level", "commit-message-fit", "commitment-strength", "comparable-fit", "completion-gate", "confirmation-match", "consent-request", "consent-scope-fit", "constraint-strength", "contact-opt-out", "content-facets", "content-freshness-signal", "context-prune", "context-role", "continuation-compare", "correction-target", "corrective-action-fit", "daw-request-kind", "deadline-risk", "defect-report-facets", "delegation-fit", "diff-hazards", "disclosure-facets", "dispute-kind", "document-role", "draft-compare", "dynamic-change", "eligibility-facets", "emotion-kind", "entity-match", "error-acknowledgment", "escalation-wording", "evaluation-mention", "evidence-conflict", "evidence-independence", "evidence-novelty", "evidence-strength", "exercise-select", "expense-category", "explanation-level", "extraction-fidelity", "fact-stability", "failure-kind", "fair-housing-wording", "feedback-actionability", "feedback-kind", "field-select", "financial-advice-signal", "finger-actions", "followup-link", "followup-timing", "forecast-confidence-wording", "format-fit", "freshness-needed", "frustration-signal", "funder-fit", "game-action", "game-phase", "goal-drift", "grounding-level", "guest-request-facets", "handoff", "handoff-completeness", "headline-fit", "incident-match", "incident-severity-wording", "injection-signal", "instruction-clarity", "instruction-conflict", "instruction-fit", "instruction-priority", "instruction-readability", "instrument-issue-kind", "instrument-report-facets", "intake-question-fit", "intent-change", "invoice-facets", "issue-impact", "issue-recurrence", "itinerary-feasibility", "job-post-facets", "length-fit", "listener-request-kind", "listing-compare", "listing-fact-consistency", "log-line-kind", "loss-cause-kind", "lyric-mood-fit", "maintenance-hazard-wording", "medication-mention", "memory-relation", "memory-scope", "memory-subject", "memory-value", "message-facets", "methods-facets", "model-route", "modulation-moment", "mood-match", "mood-request", "motivation-source", "move-explanation-fit", "narrative-consistency", "next-chord", "next-duration", "next-note", "notice-facets", "novelty-claim-level", "objection-kind", "objective-fit", "offer-terms-facets", "outcome-framing", "passage-compare", "passage-difficulty", "passage-duplicate", "passage-mood", "passage-standalone", "performance-feedback-facets", "persuasion-technique", "phrase-complete", "pii-presence", "plan-completeness", "policy-compliance", "policy-severity", "politeness-level", "postmortem-facets", "preference-kind", "priority-compare", "privacy-notice-facets", "product-match", "progress-stall", "promise-check", "proposal-facets", "qualification-evidence", "query-equivalence", "query-specificity", "question-assumption", "question-leading", "question-relevance", "reconciliation-match", "reference-resolve", "repeated-attempt", "repetition-level", "reply-template-match", "report-facets", "requirement-testability", "rerank", "resolution-check", "response-needed", "response-refusal", "result-outcome", "result-plausibility", "result-usefulness", "retrieval-needed", "retry-worthwhile", "review-comment-kind", "review-facets", "review-recommendation-kind", "review-response-fit", "rollback-signal", "root-cause-depth", "route", "route-many", "rule-compliance", "runbook-fit", "safety-incident-kind", "satisfaction-signal", "search-intent-kind", "sentiment-shift", "settlement-offer-facets", "shipment-issue-kind", "slot-fit", "sound-match", "source-applicability", "spam-signal", "stage-evidence", "step-complete", "step-progress", "style-kind", "summary-coverage", "symptom-facets", "take-turn", "task-complexity", "task-dependency", "task-duplicate", "task-overlap", "tempo-change", "tension-level", "ticket-match", "tone-check", "tool-call-gate", "tool-compare", "tool-fit", "topic-shift", "trip-purpose-kind", "troubleshooting-fit", "turn-intent", "uncertainty-expression", "urgency-signal", "verify", "wake-gate", "workaround-fit"];

package/dist/catalog/generated/names.js

@@ -25,6 +25,7 @@ export const recipeNames = [
     'bug-report-completeness',
     'buying-intent',
     'cache-match',
+    'callback-responsibility',
     'cancellation-check',
     'care-urgency-wording',
     'category-fit',
@@ -52,6 +53,7 @@ export const recipeNames = [
     'consent-request',
     'consent-scope-fit',
     'constraint-strength',
+    'contact-opt-out',
     'content-facets',
     'content-freshness-signal',
     'context-prune',
@@ -92,6 +94,7 @@ export const recipeNames = [
     'financial-advice-signal',
     'finger-actions',
     'followup-link',
+    'followup-timing',
     'forecast-confidence-wording',
     'format-fit',
     'freshness-needed',

package/dist/catalog/schema.d.ts

@@ -25,6 +25,7 @@ export declare const recipeNameSchema: z.ZodEnum<{
     "bug-report-completeness": "bug-report-completeness";
     "buying-intent": "buying-intent";
     "cache-match": "cache-match";
+    "callback-responsibility": "callback-responsibility";
     "cancellation-check": "cancellation-check";
     "care-urgency-wording": "care-urgency-wording";
     "category-fit": "category-fit";
@@ -52,6 +53,7 @@ export declare const recipeNameSchema: z.ZodEnum<{
     "consent-request": "consent-request";
     "consent-scope-fit": "consent-scope-fit";
     "constraint-strength": "constraint-strength";
+    "contact-opt-out": "contact-opt-out";
     "content-facets": "content-facets";
     "content-freshness-signal": "content-freshness-signal";
     "context-prune": "context-prune";
@@ -92,6 +94,7 @@ export declare const recipeNameSchema: z.ZodEnum<{
     "financial-advice-signal": "financial-advice-signal";
     "finger-actions": "finger-actions";
     "followup-link": "followup-link";
+    "followup-timing": "followup-timing";
     "forecast-confidence-wording": "forecast-confidence-wording";
     "format-fit": "format-fit";
     "freshness-needed": "freshness-needed";

package/dist/cli/evaluation.d.ts

@@ -0,0 +1 @@
+export declare function runEvaluationCommand(args: string[]): Promise<unknown>;

package/dist/cli/evaluation.js

@@ -0,0 +1,82 @@
+import { readFile } from 'node:fs/promises';
+import { parseArgs } from 'node:util';
+import { compare, evaluate, readEvaluationCases, readRun, replay } from '../evaluation/index.js';
+import { evaluationPolicySchema, priceSchema } from '../evaluation/schema.js';
+export async function runEvaluationCommand(args) {
+    const command = args[0];
+    if (command === 'evaluate')
+        return evaluateFromFile(args.slice(1));
+    if (command === 'replay')
+        return replayArchive(args.slice(1));
+    if (command === 'compare' && args.length === 3)
+        return compare(await readRun(args[1]), await readRun(args[2]));
+    throw new Error('Invalid evaluation command. Run jev-recipes --help for usage.');
+}
+async function evaluateFromFile(args) {
+    const { values, positionals } = parseArgs({
+        args,
+        allowPositionals: true,
+        options: {
+            cases: { type: 'string' },
+            out: { type: 'string' },
+            model: { type: 'string' },
+            concurrency: { type: 'string' },
+            'max-cases': { type: 'string' },
+            'max-requests': { type: 'string' },
+            'min-confidence': { type: 'string' },
+            split: { type: 'string' },
+            policy: { type: 'string' },
+            prices: { type: 'string' },
+        },
+    });
+    if (positionals.length !== 1 || !values.cases || !values.out)
+        throw new Error('Usage: jev-recipes evaluate <recipe> --cases <file.jsonl> --out <new-directory>');
+    if (values.policy && values['min-confidence'] !== undefined)
+        throw new Error('Use --policy or --min-confidence, not both.');
+    const policy = values.policy
+        ? evaluationPolicySchema.parse(JSON.parse(await readFile(values.policy, 'utf8')))
+        : values['min-confidence'] === undefined
+            ? {}
+            : { minConfidence: Number(values['min-confidence']) };
+    const price = values.prices
+        ? priceSchema.parse(JSON.parse(await readFile(values.prices, 'utf8')))
+        : undefined;
+    if (!process.env.TYPESAFE_API_KEY?.trim())
+        throw new Error('Set TYPESAFE_API_KEY before live evaluation. Replay and compare work offline.');
+    const run = await evaluate(positionals[0], await readEvaluationCases(values.cases), {
+        out: values.out,
+        policy,
+        ...(values.model === undefined ? {} : { model: values.model }),
+        ...(price === undefined ? {} : { price }),
+        ...(values.split === undefined ? {} : { split: values.split }),
+        ...(values.concurrency === undefined ? {} : { concurrency: Number(values.concurrency) }),
+        ...(values['max-cases'] === undefined ? {} : { maxCases: Number(values['max-cases']) }),
+        ...(values['max-requests'] === undefined
+            ? {}
+            : { maxRequests: Number(values['max-requests']) }),
+    });
+    return { archive: values.out, runId: run.runId, report: run.report };
+}
+async function replayArchive(args) {
+    const { values, positionals } = parseArgs({
+        args,
+        allowPositionals: true,
+        options: {
+            'min-confidence': { type: 'string' },
+            out: { type: 'string' },
+        },
+    });
+    if (positionals.length !== 1)
+        throw new Error('Usage: jev-recipes replay <archive> [--min-confidence <number>] [--out <new-directory>]');
+    const source = await readRun(positionals[0]);
+    const policy = values['min-confidence'] === undefined
+        ? source.policy
+        : { minConfidence: Number(values['min-confidence']) };
+    const run = await replay(source, policy, values.out);
+    return {
+        ...(values.out ? { archive: values.out } : {}),
+        sourceRun: source.runId,
+        policy: run.policy,
+        report: run.report,
+    };
+}

package/dist/cli/index.js

@@ -6,6 +6,10 @@ import { loadRecipe } from '../catalog/runner.js';
 import { listRecipes, describeRecipe, recipeCategorySchema } from '../catalog/index.js';
 import { commandArgumentsSchema, demoFixtureSchema } from './schema.js';
 async function main() {
+    if (['evaluate', 'replay', 'compare'].includes(process.argv[2] ?? '')) {
+        const { runEvaluationCommand } = await import('./evaluation.js');
+        return printJson(await runEvaluationCommand(process.argv.slice(2)));
+    }
     const command = parseCommandArguments(process.argv.slice(2));
     switch (command[0]) {
         case '--help':
@@ -40,6 +44,15 @@ function printHelp() {
   jev-recipes demo <recipe|all>         Run offline fixtures (no model calls)
   jev-recipes example <recipe>          Print example input as JSON
   jev-recipes run <recipe> <file|->      Run live with a JSON file or stdin
+  jev-recipes evaluate <recipe> --cases <file.jsonl> --out <new-directory>
+                                        Record a live evaluation and every response
+    [--concurrency 4] [--max-cases 1000] [--max-requests 1000] [--model <model>]
+    [--split development|held-out] [--min-confidence <number> | --policy <policy.json>]
+    [--prices <prices.json>]
+  jev-recipes replay <archive> [--min-confidence <number>] [--out <new-directory>]
+                                        Replay recorded responses without network calls
+  jev-recipes compare <baseline> <candidate>
+                                        Compare archives with matching data and labels
   jev-recipes --version
 
 Categories: ${recipeCategorySchema.options.join(', ')}. Quote multi-word search queries.